feat: auto --init flag, headless prompts, and prompt sanitizer (#1417)
* fix: Created 10 headless prompt files (5 workflows + 5 agents) in sdk/p… - "sdk/prompts/workflows/execute-plan.md" - "sdk/prompts/workflows/research-phase.md" - "sdk/prompts/workflows/plan-phase.md" - "sdk/prompts/workflows/verify-phase.md" - "sdk/prompts/workflows/discuss-phase.md" - "sdk/prompts/agents/gsd-executor.md" - "sdk/prompts/agents/gsd-phase-researcher.md" - "sdk/prompts/agents/gsd-planner.md" GSD-Task: S01/T02 * feat: Created prompt-sanitizer.ts, wired headless prompt loading into P… - "sdk/src/prompt-sanitizer.ts" - "sdk/src/phase-prompt.ts" - "sdk/src/gsd-tools.ts" - "sdk/src/gsd-tools.test.ts" - "sdk/src/phase-runner-types.test.ts" GSD-Task: S01/T01 * test: Added 111 unit tests covering sanitizePrompt(), headless prompt l… - "sdk/src/prompt-sanitizer.test.ts" - "sdk/src/headless-prompts.test.ts" - "sdk/src/phase-prompt.test.ts" GSD-Task: S01/T03 * feat: Wired sdkPromptsDir preference and sanitizePrompt into InitRunner… - "sdk/src/init-runner.ts" - "sdk/package.json" GSD-Task: S02/T01 * feat: add --init flag to auto command for single-command PRD-to-execution gsd-sdk auto --init @path/to/prd.md now bootstraps the project (init) then immediately runs the autonomous phase execution loop. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: add remaining headless prompt files and templates Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * test: Extended init-runner.test.ts with 7 sdkPromptsDir preference and… - "sdk/src/init-runner.test.ts" GSD-Task: S02/T03 --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
TÂCHES
@@ -198,6 +198,51 @@ describe('parseCliArgs', () => {
|
||||
|
||||
expect(result.initInput).toBeUndefined();
|
||||
});
|
||||
|
||||
// ─── Auto --init parsing ──────────────────────────────────────────────
|
||||
|
||||
it('parses auto --init with @file', () => {
|
||||
const result = parseCliArgs(['auto', '--init', '@prd.md']);
|
||||
|
||||
expect(result.command).toBe('auto');
|
||||
expect(result.init).toBe('@prd.md');
|
||||
expect(result.initInput).toBeUndefined();
|
||||
});
|
||||
|
||||
it('parses auto --init with raw text', () => {
|
||||
const result = parseCliArgs(['auto', '--init', 'build a todo app']);
|
||||
|
||||
expect(result.command).toBe('auto');
|
||||
expect(result.init).toBe('build a todo app');
|
||||
});
|
||||
|
||||
it('parses auto --init with other options', () => {
|
||||
const result = parseCliArgs([
|
||||
'auto',
|
||||
'--init', '@spec.md',
|
||||
'--project-dir', '/tmp/proj',
|
||||
'--model', 'claude-sonnet-4-6',
|
||||
'--max-budget', '25',
|
||||
]);
|
||||
|
||||
expect(result.command).toBe('auto');
|
||||
expect(result.init).toBe('@spec.md');
|
||||
expect(result.projectDir).toBe('/tmp/proj');
|
||||
expect(result.model).toBe('claude-sonnet-4-6');
|
||||
expect(result.maxBudget).toBe(25);
|
||||
});
|
||||
|
||||
it('init is undefined when --init not provided', () => {
|
||||
const result = parseCliArgs(['auto']);
|
||||
|
||||
expect(result.init).toBeUndefined();
|
||||
});
|
||||
|
||||
it('init is undefined for non-auto commands', () => {
|
||||
const result = parseCliArgs(['run', 'build auth']);
|
||||
|
||||
expect(result.init).toBeUndefined();
|
||||
});
|
||||
});
|
||||
|
||||
// ─── resolveInitInput tests ──────────────────────────────────────────────────
|
||||
@@ -219,6 +264,7 @@ describe('resolveInitInput', () => {
|
||||
command: 'init',
|
||||
prompt: undefined,
|
||||
initInput: undefined,
|
||||
init: undefined,
|
||||
projectDir: tmpDir,
|
||||
wsPort: undefined,
|
||||
model: undefined,
|
||||
@@ -307,4 +353,9 @@ describe('USAGE', () => {
|
||||
it('describes auto as autonomous lifecycle', () => {
|
||||
expect(USAGE).toMatch(/auto\s+.*autonomous/i);
|
||||
});
|
||||
|
||||
it('documents --init option', () => {
|
||||
expect(USAGE).toContain('--init');
|
||||
expect(USAGE).toContain('Bootstrap from a PRD');
|
||||
});
|
||||
});
|
||||
|
||||
@@ -23,6 +23,8 @@ export interface ParsedCliArgs {
|
||||
prompt: string | undefined;
|
||||
/** For 'init' command: the raw input source (@file, text, or undefined for stdin). */
|
||||
initInput: string | undefined;
|
||||
/** For 'auto --init': bootstrap from a PRD before running the autonomous loop. */
|
||||
init: string | undefined;
|
||||
projectDir: string;
|
||||
wsPort: number | undefined;
|
||||
model: string | undefined;
|
||||
@@ -43,6 +45,7 @@ export function parseCliArgs(argv: string[]): ParsedCliArgs {
|
||||
'ws-port': { type: 'string' },
|
||||
model: { type: 'string' },
|
||||
'max-budget': { type: 'string' },
|
||||
init: { type: 'string' },
|
||||
help: { type: 'boolean', short: 'h', default: false },
|
||||
version: { type: 'boolean', short: 'v', default: false },
|
||||
},
|
||||
@@ -61,6 +64,7 @@ export function parseCliArgs(argv: string[]): ParsedCliArgs {
|
||||
command,
|
||||
prompt,
|
||||
initInput,
|
||||
init: values.init as string | undefined,
|
||||
projectDir: values['project-dir'] as string,
|
||||
wsPort: values['ws-port'] ? Number(values['ws-port']) : undefined,
|
||||
model: values.model as string | undefined,
|
||||
@@ -85,6 +89,8 @@ Commands:
|
||||
(empty) Read from stdin
|
||||
|
||||
Options:
|
||||
--init <input> Bootstrap from a PRD before running (auto only)
|
||||
Accepts @path/to/prd.md or "description text"
|
||||
--project-dir <dir> Project directory (default: cwd)
|
||||
--ws-port <port> Enable WebSocket transport on <port>
|
||||
--model <model> Override LLM model
|
||||
@@ -306,6 +312,47 @@ export async function main(argv: string[] = process.argv.slice(2)): Promise<void
|
||||
}
|
||||
|
||||
try {
|
||||
// If --init provided, bootstrap project first
|
||||
if (args.init) {
|
||||
const initInput = await resolveInitInput({
|
||||
...args,
|
||||
command: 'init',
|
||||
initInput: args.init,
|
||||
});
|
||||
|
||||
console.log(`[auto] Bootstrapping project from --init (${initInput.length} chars)`);
|
||||
|
||||
const tools = gsd.createTools();
|
||||
const runner = new InitRunner({
|
||||
projectDir: args.projectDir,
|
||||
tools,
|
||||
eventStream: gsd.eventStream,
|
||||
config: {
|
||||
maxBudgetPerSession: args.maxBudget,
|
||||
orchestratorModel: args.model,
|
||||
},
|
||||
});
|
||||
|
||||
const initResult = await runner.run(initInput);
|
||||
|
||||
const initStatus = initResult.success ? 'SUCCESS' : 'FAILED';
|
||||
const stepCount = initResult.steps.length;
|
||||
const passedSteps = initResult.steps.filter(s => s.success).length;
|
||||
const initCost = initResult.totalCostUsd.toFixed(2);
|
||||
const initDuration = (initResult.totalDurationMs / 1000).toFixed(1);
|
||||
console.log(`[init ${initStatus}] ${passedSteps}/${stepCount} steps, $${initCost}, ${initDuration}s`);
|
||||
|
||||
if (!initResult.success) {
|
||||
for (const step of initResult.steps) {
|
||||
if (!step.success && step.error) {
|
||||
console.error(` ✗ ${step.step}: ${step.error}`);
|
||||
}
|
||||
}
|
||||
process.exitCode = 1;
|
||||
return;
|
||||
}
|
||||
}
|
||||
|
||||
const result = await gsd.run('');
|
||||
|
||||
// Final summary
|
||||
|
||||
@@ -287,7 +287,7 @@ describe('GSDTools', () => {
|
||||
'init-new-project.cjs',
|
||||
`
|
||||
const args = process.argv.slice(2);
|
||||
if (args[0] === 'init' && args[1] === 'new-project' && args.includes('--raw')) {
|
||||
if (args[0] === 'init' && args[1] === 'new-project') {
|
||||
process.stdout.write(JSON.stringify(${JSON.stringify(mockResult)}));
|
||||
} else {
|
||||
process.stderr.write('unexpected args: ' + args.join(' '));
|
||||
|
||||
@@ -51,11 +51,10 @@ export class GSDTools {
|
||||
|
||||
/**
|
||||
* Execute a gsd-tools command and return parsed JSON output.
|
||||
* Appends `--raw` to get machine-readable JSON output.
|
||||
* Handles the `@file:` prefix pattern for large results.
|
||||
*/
|
||||
async exec(command: string, args: string[] = []): Promise<unknown> {
|
||||
const fullArgs = [this.gsdToolsPath, command, ...args, '--raw'];
|
||||
const fullArgs = [this.gsdToolsPath, command, ...args];
|
||||
|
||||
return new Promise<unknown>((resolve, reject) => {
|
||||
const child = execFile(
|
||||
|
||||
159
sdk/src/headless-prompts.test.ts
Normal file
159
sdk/src/headless-prompts.test.ts
Normal file
@@ -0,0 +1,159 @@
|
||||
/**
|
||||
* Contract test: all headless prompt files in sdk/prompts/ must contain
|
||||
* zero instances of blocked interactive patterns.
|
||||
*
|
||||
* This prevents regression — any new prompt file or edit that reintroduces
|
||||
* interactive mechanics will fail this test.
|
||||
*/
|
||||
import { describe, it, expect } from 'vitest';
|
||||
import { readFile } from 'node:fs/promises';
|
||||
import { join, dirname } from 'node:path';
|
||||
import { fileURLToPath } from 'node:url';
|
||||
import { readdirSync } from 'node:fs';
|
||||
|
||||
// ─── Paths ───────────────────────────────────────────────────────────────────
|
||||
|
||||
const __dirname = dirname(fileURLToPath(import.meta.url));
|
||||
const promptsDir = join(__dirname, '..', 'prompts');
|
||||
const workflowsDir = join(promptsDir, 'workflows');
|
||||
const agentsDir = join(promptsDir, 'agents');
|
||||
|
||||
// ─── Blocked patterns ────────────────────────────────────────────────────────
|
||||
|
||||
/**
|
||||
* Patterns that MUST NOT appear in headless prompts.
|
||||
* Each entry: [label for reporting, regex].
|
||||
*/
|
||||
const BLOCKED_PATTERNS: Array<[string, RegExp]> = [
|
||||
['AskUserQuestion', /AskUserQuestion\s*\(/],
|
||||
['SlashCommand', /SlashCommand\s*\(/],
|
||||
['/gsd: command', /\/gsd:\S+/],
|
||||
['@file: reference', /@file:\S+/],
|
||||
['STOP + wait directive', /\bSTOP\b\s+(?:and\s+)?(?:wait|ask)/i],
|
||||
['bare STOP directive', /^\s*STOP\s*[.!]?\s*$/m],
|
||||
['wait for user', /\bwait\s+for\s+(?:the\s+)?user\b/i],
|
||||
['ask the user', /\bask\s+the\s+user\b/i],
|
||||
];
|
||||
|
||||
// ─── Expected files ──────────────────────────────────────────────────────────
|
||||
|
||||
const EXPECTED_WORKFLOWS = [
|
||||
'execute-plan.md',
|
||||
'research-phase.md',
|
||||
'plan-phase.md',
|
||||
'verify-phase.md',
|
||||
'discuss-phase.md',
|
||||
];
|
||||
|
||||
const EXPECTED_AGENTS = [
|
||||
'gsd-executor.md',
|
||||
'gsd-phase-researcher.md',
|
||||
'gsd-planner.md',
|
||||
'gsd-verifier.md',
|
||||
'gsd-plan-checker.md',
|
||||
'gsd-project-researcher.md',
|
||||
'gsd-research-synthesizer.md',
|
||||
'gsd-roadmapper.md',
|
||||
];
|
||||
|
||||
const templatesDir = join(promptsDir, 'templates');
|
||||
const researchTemplatesDir = join(templatesDir, 'research-project');
|
||||
|
||||
const EXPECTED_TEMPLATES = [
|
||||
'project.md',
|
||||
'requirements.md',
|
||||
'roadmap.md',
|
||||
'state.md',
|
||||
];
|
||||
|
||||
const EXPECTED_RESEARCH_TEMPLATES = [
|
||||
'ARCHITECTURE.md',
|
||||
'FEATURES.md',
|
||||
'PITFALLS.md',
|
||||
'STACK.md',
|
||||
'SUMMARY.md',
|
||||
];
|
||||
|
||||
// ─── Tests ───────────────────────────────────────────────────────────────────
|
||||
|
||||
describe('headless prompt contract', () => {
|
||||
describe('file inventory', () => {
|
||||
it('has all expected workflow files', () => {
|
||||
const actual = readdirSync(workflowsDir).sort();
|
||||
expect(actual).toEqual(EXPECTED_WORKFLOWS.sort());
|
||||
});
|
||||
|
||||
it('has all expected agent files', () => {
|
||||
const actual = readdirSync(agentsDir).sort();
|
||||
expect(actual).toEqual(EXPECTED_AGENTS.sort());
|
||||
});
|
||||
});
|
||||
|
||||
describe('zero interactive patterns in workflow prompts', () => {
|
||||
for (const filename of EXPECTED_WORKFLOWS) {
|
||||
describe(filename, () => {
|
||||
for (const [label, pattern] of BLOCKED_PATTERNS) {
|
||||
it(`contains no ${label}`, async () => {
|
||||
const content = await readFile(join(workflowsDir, filename), 'utf-8');
|
||||
const matches = content.match(new RegExp(pattern.source, pattern.flags + 'g'));
|
||||
expect(matches, `Found ${label} in ${filename}: ${matches?.join(', ')}`).toBeNull();
|
||||
});
|
||||
}
|
||||
});
|
||||
}
|
||||
});
|
||||
|
||||
describe('zero interactive patterns in agent prompts', () => {
|
||||
for (const filename of EXPECTED_AGENTS) {
|
||||
describe(filename, () => {
|
||||
for (const [label, pattern] of BLOCKED_PATTERNS) {
|
||||
it(`contains no ${label}`, async () => {
|
||||
const content = await readFile(join(agentsDir, filename), 'utf-8');
|
||||
const matches = content.match(new RegExp(pattern.source, pattern.flags + 'g'));
|
||||
expect(matches, `Found ${label} in ${filename}: ${matches?.join(', ')}`).toBeNull();
|
||||
});
|
||||
}
|
||||
});
|
||||
}
|
||||
});
|
||||
|
||||
describe('template file inventory', () => {
|
||||
it('has all expected top-level template files', () => {
|
||||
const actual = readdirSync(templatesDir).filter(f => f.endsWith('.md')).sort();
|
||||
expect(actual).toEqual(EXPECTED_TEMPLATES.sort());
|
||||
});
|
||||
|
||||
it('has all expected research-project template files', () => {
|
||||
const actual = readdirSync(researchTemplatesDir).sort();
|
||||
expect(actual).toEqual(EXPECTED_RESEARCH_TEMPLATES.sort());
|
||||
});
|
||||
});
|
||||
|
||||
describe('zero interactive patterns in template prompts', () => {
|
||||
for (const filename of EXPECTED_TEMPLATES) {
|
||||
describe(filename, () => {
|
||||
for (const [label, pattern] of BLOCKED_PATTERNS) {
|
||||
it(`contains no ${label}`, async () => {
|
||||
const content = await readFile(join(templatesDir, filename), 'utf-8');
|
||||
const matches = content.match(new RegExp(pattern.source, pattern.flags + 'g'));
|
||||
expect(matches, `Found ${label} in ${filename}: ${matches?.join(', ')}`).toBeNull();
|
||||
});
|
||||
}
|
||||
});
|
||||
}
|
||||
});
|
||||
|
||||
describe('zero interactive patterns in research-project templates', () => {
|
||||
for (const filename of EXPECTED_RESEARCH_TEMPLATES) {
|
||||
describe(filename, () => {
|
||||
for (const [label, pattern] of BLOCKED_PATTERNS) {
|
||||
it(`contains no ${label}`, async () => {
|
||||
const content = await readFile(join(researchTemplatesDir, filename), 'utf-8');
|
||||
const matches = content.match(new RegExp(pattern.source, pattern.flags + 'g'));
|
||||
expect(matches, `Found ${label} in ${filename}: ${matches?.join(', ')}`).toBeNull();
|
||||
});
|
||||
}
|
||||
});
|
||||
}
|
||||
});
|
||||
});
|
||||
@@ -6,7 +6,7 @@
|
||||
*
|
||||
* @example
|
||||
* ```typescript
|
||||
* import { GSD } from '@gsd/sdk';
|
||||
* import { GSD } from '@gsd-build/sdk';
|
||||
*
|
||||
* const gsd = new GSD({ projectDir: '/path/to/project' });
|
||||
* const result = await gsd.executePlan('.planning/phases/01-auth/01-auth-01-PLAN.md');
|
||||
|
||||
@@ -560,4 +560,224 @@ describe('InitRunner', () => {
|
||||
// 1 PROJECT.md + 4 research + 1 synthesis + 1 requirements + 1 roadmap = 8
|
||||
expect(mockRunSession).toHaveBeenCalledTimes(8);
|
||||
});
|
||||
|
||||
// ─── Headless prompt loading (sdkPromptsDir preference) ──────────────────
|
||||
|
||||
describe('sdkPromptsDir preference and sanitizer integration', () => {
|
||||
let sdkPromptsDir: string;
|
||||
|
||||
beforeEach(async () => {
|
||||
// Create a temp SDK prompts directory with test fixtures
|
||||
sdkPromptsDir = join(tmpDir, 'sdk-prompts');
|
||||
await mkdir(join(sdkPromptsDir, 'templates', 'research-project'), { recursive: true });
|
||||
await mkdir(join(sdkPromptsDir, 'agents'), { recursive: true });
|
||||
|
||||
// Write headless templates (with known marker text for assertion)
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'project.md'),
|
||||
'# PROJECT Template\nSDK_HEADLESS_MARKER_PROJECT\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'requirements.md'),
|
||||
'# REQUIREMENTS Template\nSDK_HEADLESS_MARKER_REQUIREMENTS\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'roadmap.md'),
|
||||
'# ROADMAP Template\nSDK_HEADLESS_MARKER_ROADMAP\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'state.md'),
|
||||
'# STATE Template\nSDK_HEADLESS_MARKER_STATE\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'research-project', 'STACK.md'),
|
||||
'# STACK Template\nSDK_HEADLESS_MARKER_STACK\n',
|
||||
);
|
||||
|
||||
// Write headless agents (with known marker text)
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'agents', 'gsd-project-researcher.md'),
|
||||
'# Project Researcher Agent\nSDK_HEADLESS_MARKER_RESEARCHER\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'agents', 'gsd-research-synthesizer.md'),
|
||||
'# Research Synthesizer Agent\nSDK_HEADLESS_MARKER_SYNTHESIZER\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'agents', 'gsd-roadmapper.md'),
|
||||
'# Roadmapper Agent\nSDK_HEADLESS_MARKER_ROADMAPPER\n',
|
||||
);
|
||||
});
|
||||
|
||||
function createRunnerWithSdkPrompts(
|
||||
toolsOverrides: Record<string, unknown> = {},
|
||||
configOverrides?: Partial<InitRunnerDeps['config']>,
|
||||
) {
|
||||
const tools = makeTools(toolsOverrides);
|
||||
const eventStream = makeEventStream();
|
||||
const runner = new InitRunner({
|
||||
projectDir: tmpDir,
|
||||
tools,
|
||||
eventStream,
|
||||
config: configOverrides as any,
|
||||
sdkPromptsDir,
|
||||
});
|
||||
return { runner, tools, eventStream, events: eventStream.events as GSDEvent[] };
|
||||
}
|
||||
|
||||
it('readGSDFile prefers sdk/prompts/ template over GSD-1 path', async () => {
|
||||
const { runner } = createRunnerWithSdkPrompts();
|
||||
|
||||
await runner.run('build a todo app');
|
||||
|
||||
// The first session call is buildProjectPrompt → reads templates/project.md
|
||||
const projectPrompt = mockRunSession.mock.calls[0]![0] as string;
|
||||
expect(projectPrompt).toContain('SDK_HEADLESS_MARKER_PROJECT');
|
||||
});
|
||||
|
||||
it('readAgentFile prefers sdk/prompts/agents/ over GSD-1 path', async () => {
|
||||
const { runner } = createRunnerWithSdkPrompts();
|
||||
|
||||
await runner.run('build a todo app');
|
||||
|
||||
// Research calls (indices 1-4) use gsd-project-researcher.md agent def
|
||||
const researchPrompt = mockRunSession.mock.calls[1]![0] as string;
|
||||
expect(researchPrompt).toContain('SDK_HEADLESS_MARKER_RESEARCHER');
|
||||
});
|
||||
|
||||
it('readGSDFile falls back to GSD-1 when sdk/prompts/ file does not exist', async () => {
|
||||
// Create an empty sdkPromptsDir — no templates at all
|
||||
const emptySdkDir = join(tmpDir, 'empty-sdk-prompts');
|
||||
await mkdir(join(emptySdkDir, 'templates'), { recursive: true });
|
||||
await mkdir(join(emptySdkDir, 'agents'), { recursive: true });
|
||||
|
||||
const tools = makeTools();
|
||||
const eventStream = makeEventStream();
|
||||
const runner = new InitRunner({
|
||||
projectDir: tmpDir,
|
||||
tools,
|
||||
eventStream,
|
||||
sdkPromptsDir: emptySdkDir,
|
||||
});
|
||||
|
||||
await runner.run('build a todo app');
|
||||
|
||||
// buildProjectPrompt reads templates/project.md — not found in empty dir,
|
||||
// falls through to GSD-1 path. If GSD-1 also missing, gets placeholder.
|
||||
const projectPrompt = mockRunSession.mock.calls[0]![0] as string;
|
||||
|
||||
// Should NOT contain our marker (since empty dir was used)
|
||||
expect(projectPrompt).not.toContain('SDK_HEADLESS_MARKER_PROJECT');
|
||||
// Should still contain the PROJECT.md synthesis instruction (from the prompt builder)
|
||||
expect(projectPrompt).toContain('PROJECT.md');
|
||||
});
|
||||
|
||||
it('readAgentFile falls back to GSD-1 when sdk/prompts/agents/ file does not exist', async () => {
|
||||
// Empty sdkPromptsDir — no agent files
|
||||
const emptySdkDir = join(tmpDir, 'empty-sdk-agents');
|
||||
await mkdir(join(emptySdkDir, 'templates', 'research-project'), { recursive: true });
|
||||
await mkdir(join(emptySdkDir, 'agents'), { recursive: true });
|
||||
|
||||
// Write templates so we get past buildProjectPrompt
|
||||
await writeFile(join(emptySdkDir, 'templates', 'project.md'), '# project\n');
|
||||
await writeFile(join(emptySdkDir, 'templates', 'research-project', 'STACK.md'), '# stack\n');
|
||||
await writeFile(join(emptySdkDir, 'templates', 'research-project', 'FEATURES.md'), '# features\n');
|
||||
await writeFile(join(emptySdkDir, 'templates', 'research-project', 'ARCHITECTURE.md'), '# arch\n');
|
||||
await writeFile(join(emptySdkDir, 'templates', 'research-project', 'PITFALLS.md'), '# pitfalls\n');
|
||||
|
||||
const tools = makeTools();
|
||||
const eventStream = makeEventStream();
|
||||
const runner = new InitRunner({
|
||||
projectDir: tmpDir,
|
||||
tools,
|
||||
eventStream,
|
||||
sdkPromptsDir: emptySdkDir,
|
||||
});
|
||||
|
||||
await runner.run('build a todo app');
|
||||
|
||||
// Research prompt uses agent def — not in empty agents dir, falls to GSD-1
|
||||
const researchPrompt = mockRunSession.mock.calls[1]![0] as string;
|
||||
// Should NOT contain our marker
|
||||
expect(researchPrompt).not.toContain('SDK_HEADLESS_MARKER_RESEARCHER');
|
||||
// Should still have the "researching the" instruction
|
||||
expect(researchPrompt).toContain('You are researching the');
|
||||
});
|
||||
|
||||
it('buildProjectPrompt output passes through sanitizePrompt (no /gsd: patterns)', async () => {
|
||||
// Write a template that contains an interactive pattern
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'project.md'),
|
||||
'# PROJECT Template\nRun /gsd:map-codebase to analyze.\nSDK_HEADLESS_MARKER_PROJECT\n',
|
||||
);
|
||||
|
||||
const { runner } = createRunnerWithSdkPrompts();
|
||||
await runner.run('build a todo app');
|
||||
|
||||
const projectPrompt = mockRunSession.mock.calls[0]![0] as string;
|
||||
// sanitizePrompt should have stripped the /gsd: line
|
||||
expect(projectPrompt).not.toMatch(/\/gsd:\S+/);
|
||||
// But the marker should still be there
|
||||
expect(projectPrompt).toContain('SDK_HEADLESS_MARKER_PROJECT');
|
||||
});
|
||||
|
||||
it('buildResearchPrompt output passes through sanitizePrompt (no /gsd: patterns)', async () => {
|
||||
// Write an agent def that contains interactive patterns
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'agents', 'gsd-project-researcher.md'),
|
||||
'# Researcher Agent\nSpawn /gsd:something for analysis.\nSDK_HEADLESS_MARKER_RESEARCHER\n',
|
||||
);
|
||||
|
||||
const { runner } = createRunnerWithSdkPrompts();
|
||||
await runner.run('build a todo app');
|
||||
|
||||
const researchPrompt = mockRunSession.mock.calls[1]![0] as string;
|
||||
// sanitizePrompt should have stripped the /gsd: line
|
||||
expect(researchPrompt).not.toMatch(/\/gsd:\S+/);
|
||||
// Marker should still be present
|
||||
expect(researchPrompt).toContain('SDK_HEADLESS_MARKER_RESEARCHER');
|
||||
});
|
||||
|
||||
it('buildRoadmapPrompt output passes through sanitizePrompt (no /gsd: patterns)', async () => {
|
||||
// Write agent and templates with interactive patterns
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'agents', 'gsd-roadmapper.md'),
|
||||
'# Roadmapper Agent\nUse /gsd:execute to run.\nSDK_HEADLESS_MARKER_ROADMAPPER\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'roadmap.md'),
|
||||
'# ROADMAP Template\nRun /gsd:check-progress.\nSDK_HEADLESS_MARKER_ROADMAP\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'state.md'),
|
||||
'# STATE Template\nUse /gsd:add-todo for tracking.\nSDK_HEADLESS_MARKER_STATE\n',
|
||||
);
|
||||
|
||||
// Also need research templates and synth agent for earlier steps
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'research-project', 'FEATURES.md'), '# features\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'research-project', 'ARCHITECTURE.md'), '# arch\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'research-project', 'PITFALLS.md'), '# pitfalls\n',
|
||||
);
|
||||
await writeFile(
|
||||
join(sdkPromptsDir, 'templates', 'research-project', 'SUMMARY.md'), '# summary\n',
|
||||
);
|
||||
|
||||
const { runner } = createRunnerWithSdkPrompts();
|
||||
await runner.run('build a todo app');
|
||||
|
||||
// Roadmap prompt is the last session call (index 7)
|
||||
const roadmapPrompt = mockRunSession.mock.calls[7]![0] as string;
|
||||
// sanitizePrompt should have stripped all /gsd: patterns
|
||||
expect(roadmapPrompt).not.toMatch(/\/gsd:\S+/);
|
||||
// Markers from templates should still be present
|
||||
expect(roadmapPrompt).toContain('SDK_HEADLESS_MARKER_ROADMAPPER');
|
||||
expect(roadmapPrompt).toContain('SDK_HEADLESS_MARKER_ROADMAP');
|
||||
expect(roadmapPrompt).toContain('SDK_HEADLESS_MARKER_STATE');
|
||||
});
|
||||
});
|
||||
});
|
||||
|
||||
@@ -10,6 +10,7 @@
|
||||
|
||||
import { readFile, writeFile, mkdir } from 'node:fs/promises';
|
||||
import { join } from 'node:path';
|
||||
import { fileURLToPath } from 'node:url';
|
||||
import { homedir } from 'node:os';
|
||||
import { execFile } from 'node:child_process';
|
||||
|
||||
@@ -31,6 +32,7 @@ import type { GSDTools } from './gsd-tools.js';
|
||||
import type { GSDEventStream } from './event-stream.js';
|
||||
import { loadConfig } from './config.js';
|
||||
import { runPhaseStepSession } from './session-runner.js';
|
||||
import { sanitizePrompt } from './prompt-sanitizer.js';
|
||||
|
||||
// ─── Constants ───────────────────────────────────────────────────────────────
|
||||
|
||||
@@ -68,6 +70,8 @@ export interface InitRunnerDeps {
|
||||
tools: GSDTools;
|
||||
eventStream: GSDEventStream;
|
||||
config?: Partial<InitConfig>;
|
||||
/** Override for SDK prompts directory. Defaults to package-relative sdk/prompts/. */
|
||||
sdkPromptsDir?: string;
|
||||
}
|
||||
|
||||
export class InitRunner {
|
||||
@@ -76,6 +80,7 @@ export class InitRunner {
|
||||
private readonly eventStream: GSDEventStream;
|
||||
private readonly config: InitConfig;
|
||||
private readonly sessionId: string;
|
||||
private readonly sdkPromptsDir: string;
|
||||
|
||||
constructor(deps: InitRunnerDeps) {
|
||||
this.projectDir = deps.projectDir;
|
||||
@@ -88,6 +93,10 @@ export class InitRunner {
|
||||
orchestratorModel: deps.config?.orchestratorModel,
|
||||
};
|
||||
this.sessionId = `init-${Date.now()}`;
|
||||
// SDK prompts dir: explicit override → package-relative default via import.meta.url
|
||||
this.sdkPromptsDir =
|
||||
deps.sdkPromptsDir ??
|
||||
join(fileURLToPath(new URL('.', import.meta.url)), '..', 'prompts');
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -375,7 +384,7 @@ export class InitRunner {
|
||||
private async buildProjectPrompt(input: string): Promise<string> {
|
||||
const template = await this.readGSDFile('templates/project.md');
|
||||
|
||||
return [
|
||||
return sanitizePrompt([
|
||||
'You are creating the PROJECT.md for a new software project.',
|
||||
'Write .planning/PROJECT.md based on the template structure below and the user\'s project description.',
|
||||
'',
|
||||
@@ -389,7 +398,7 @@ export class InitRunner {
|
||||
'',
|
||||
'Write the file to .planning/PROJECT.md. Follow the template structure but fill in with real content derived from the user input.',
|
||||
'Be specific and opinionated — make decisions, don\'t list options.',
|
||||
].join('\n');
|
||||
].join('\n'));
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -415,7 +424,7 @@ export class InitRunner {
|
||||
projectContent = input;
|
||||
}
|
||||
|
||||
return [
|
||||
return sanitizePrompt([
|
||||
'<agent_definition>',
|
||||
agentDef,
|
||||
'</agent_definition>',
|
||||
@@ -437,7 +446,7 @@ export class InitRunner {
|
||||
'',
|
||||
`Write .planning/research/${researchType}.md following the template structure.`,
|
||||
'Be comprehensive but opinionated. "Use X because Y" not "Options are X, Y, Z."',
|
||||
].join('\n');
|
||||
].join('\n'));
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -460,7 +469,7 @@ export class InitRunner {
|
||||
}
|
||||
}
|
||||
|
||||
return [
|
||||
return sanitizePrompt([
|
||||
'<agent_definition>',
|
||||
agentDef,
|
||||
'</agent_definition>',
|
||||
@@ -482,7 +491,7 @@ export class InitRunner {
|
||||
'',
|
||||
'Write .planning/research/SUMMARY.md synthesizing all research findings.',
|
||||
'Also commit all research files: git add .planning/research/ && git commit.',
|
||||
].join('\n');
|
||||
].join('\n'));
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -511,7 +520,7 @@ export class InitRunner {
|
||||
// Research may have partially failed
|
||||
}
|
||||
|
||||
return [
|
||||
return sanitizePrompt([
|
||||
'You are generating REQUIREMENTS.md for this project.',
|
||||
'Derive requirements from the PROJECT.md and research outputs.',
|
||||
'Auto-include all table-stakes requirements (auth, error handling, logging, etc.).',
|
||||
@@ -530,7 +539,7 @@ export class InitRunner {
|
||||
'',
|
||||
'Write .planning/REQUIREMENTS.md following the template structure.',
|
||||
'Every requirement must be testable and specific. No vague aspirations.',
|
||||
].join('\n');
|
||||
].join('\n'));
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -559,7 +568,7 @@ export class InitRunner {
|
||||
}
|
||||
}
|
||||
|
||||
return [
|
||||
return sanitizePrompt([
|
||||
'<agent_definition>',
|
||||
agentDef,
|
||||
'</agent_definition>',
|
||||
@@ -581,7 +590,7 @@ export class InitRunner {
|
||||
'Create .planning/ROADMAP.md and .planning/STATE.md.',
|
||||
'ROADMAP.md: Transform requirements into phases. Every v1 requirement maps to exactly one phase.',
|
||||
'STATE.md: Initialize project state tracking.',
|
||||
].join('\n');
|
||||
].join('\n'));
|
||||
}
|
||||
|
||||
// ─── Session execution ─────────────────────────────────────────────────────
|
||||
@@ -610,9 +619,20 @@ export class InitRunner {
|
||||
// ─── File reading helpers ──────────────────────────────────────────────────
|
||||
|
||||
/**
|
||||
* Read a file from the GSD templates directory (~/.claude/get-shit-done/).
|
||||
* Read a file from the GSD templates directory.
|
||||
* Tries sdk/prompts/{relativePath} first (headless versions), then
|
||||
* falls back to GSD-1 originals (~/.claude/get-shit-done/).
|
||||
*/
|
||||
private async readGSDFile(relativePath: string): Promise<string> {
|
||||
// Try SDK prompts dir first (headless versions)
|
||||
const sdkPath = join(this.sdkPromptsDir, relativePath);
|
||||
try {
|
||||
return await readFile(sdkPath, 'utf-8');
|
||||
} catch {
|
||||
// Not in sdk/prompts/, fall through to GSD-1 originals
|
||||
}
|
||||
|
||||
// Fall back to GSD-1 originals
|
||||
const fullPath = join(GSD_TEMPLATES_DIR, '..', relativePath);
|
||||
try {
|
||||
return await readFile(fullPath, 'utf-8');
|
||||
@@ -623,9 +643,20 @@ export class InitRunner {
|
||||
}
|
||||
|
||||
/**
|
||||
* Read an agent definition from ~/.claude/agents/.
|
||||
* Read an agent definition.
|
||||
* Tries sdk/prompts/agents/{filename} first (headless versions), then
|
||||
* falls back to GSD-1 originals (~/.claude/agents/).
|
||||
*/
|
||||
private async readAgentFile(filename: string): Promise<string> {
|
||||
// Try SDK prompts dir first (headless versions)
|
||||
const sdkPath = join(this.sdkPromptsDir, 'agents', filename);
|
||||
try {
|
||||
return await readFile(sdkPath, 'utf-8');
|
||||
} catch {
|
||||
// Not in sdk/prompts/, fall through to GSD-1 originals
|
||||
}
|
||||
|
||||
// Fall back to GSD-1 originals
|
||||
const fullPath = join(GSD_AGENTS_DIR, filename);
|
||||
try {
|
||||
return await readFile(fullPath, 'utf-8');
|
||||
|
||||
@@ -116,9 +116,12 @@ describe('PromptFactory', () => {
|
||||
});
|
||||
|
||||
function makeFactory(): PromptFactory {
|
||||
// sdkPromptsDir points to a non-existent temp subdir so real sdk/prompts/ files
|
||||
// don't interfere — tests control exactly which files exist on disk.
|
||||
return new PromptFactory({
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
sdkPromptsDir: join(tempDir, 'sdk-prompts-does-not-exist'),
|
||||
});
|
||||
}
|
||||
|
||||
@@ -365,6 +368,7 @@ describe('PromptFactory', () => {
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
projectAgentsDir,
|
||||
sdkPromptsDir: join(tempDir, 'sdk-prompts-does-not-exist'),
|
||||
});
|
||||
|
||||
const content = await factory.loadAgentDef(PhaseType.Execute);
|
||||
@@ -381,12 +385,138 @@ describe('PromptFactory', () => {
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
projectAgentsDir,
|
||||
sdkPromptsDir: join(tempDir, 'sdk-prompts-does-not-exist'),
|
||||
});
|
||||
|
||||
const content = await factory.loadAgentDef(PhaseType.Execute);
|
||||
expect(content).toBe('user agent');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── Headless prompt loading ─────────────────────────────────────────────
|
||||
|
||||
describe('headless prompt loading', () => {
|
||||
it('loadWorkflowFile prefers sdkPromptsDir over GSD-1 workflowsDir', async () => {
|
||||
const sdkDir = join(tempDir, 'sdk-prompts');
|
||||
await mkdir(join(sdkDir, 'workflows'), { recursive: true });
|
||||
|
||||
// Write both: GSD-1 original and SDK headless version
|
||||
await writeFile(join(workflowsDir, 'research-phase.md'), 'GSD-1 original');
|
||||
await writeFile(join(sdkDir, 'workflows', 'research-phase.md'), 'SDK headless version');
|
||||
|
||||
const factory = new PromptFactory({
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
sdkPromptsDir: sdkDir,
|
||||
});
|
||||
|
||||
const content = await factory.loadWorkflowFile(PhaseType.Research);
|
||||
expect(content).toBe('SDK headless version');
|
||||
});
|
||||
|
||||
it('loadWorkflowFile falls back to GSD-1 when sdkPromptsDir file missing', async () => {
|
||||
const sdkDir = join(tempDir, 'sdk-prompts');
|
||||
await mkdir(join(sdkDir, 'workflows'), { recursive: true });
|
||||
|
||||
// Only GSD-1 original exists, no SDK version
|
||||
await writeFile(join(workflowsDir, 'research-phase.md'), 'GSD-1 original');
|
||||
|
||||
const factory = new PromptFactory({
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
sdkPromptsDir: sdkDir,
|
||||
});
|
||||
|
||||
const content = await factory.loadWorkflowFile(PhaseType.Research);
|
||||
expect(content).toBe('GSD-1 original');
|
||||
});
|
||||
|
||||
it('loadAgentDef prefers sdkPromptsDir over user agents dir', async () => {
|
||||
const sdkDir = join(tempDir, 'sdk-prompts');
|
||||
await mkdir(join(sdkDir, 'agents'), { recursive: true });
|
||||
|
||||
// Write both: user agent and SDK headless agent
|
||||
await writeFile(join(agentsDir, 'gsd-executor.md'), 'user agent');
|
||||
await writeFile(join(sdkDir, 'agents', 'gsd-executor.md'), 'SDK headless agent');
|
||||
|
||||
const factory = new PromptFactory({
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
sdkPromptsDir: sdkDir,
|
||||
});
|
||||
|
||||
const content = await factory.loadAgentDef(PhaseType.Execute);
|
||||
expect(content).toBe('SDK headless agent');
|
||||
});
|
||||
|
||||
it('loadAgentDef falls back to user agents when sdkPromptsDir file missing', async () => {
|
||||
const sdkDir = join(tempDir, 'sdk-prompts');
|
||||
await mkdir(join(sdkDir, 'agents'), { recursive: true });
|
||||
|
||||
// Only user agent exists, no SDK version
|
||||
await writeFile(join(agentsDir, 'gsd-executor.md'), 'user agent');
|
||||
|
||||
const factory = new PromptFactory({
|
||||
gsdInstallDir: tempDir,
|
||||
agentsDir,
|
||||
sdkPromptsDir: sdkDir,
|
||||
});
|
||||
|
||||
const content = await factory.loadAgentDef(PhaseType.Execute);
|
||||
expect(content).toBe('user agent');
|
||||
});
|
||||
|
||||
it('buildPrompt sanitizes interactive patterns from output', async () => {
|
||||
// Use separate lines so non-interactive content survives stripping
|
||||
await writeFile(
|
||||
join(workflowsDir, 'research-phase.md'),
|
||||
makeWorkflowContent('Research the codebase thoroughly.', [
|
||||
'Gather data from the project.\nAskUserQuestion("what?")\nAnalyze findings.',
|
||||
'Run the analysis.\n/gsd:analyze --deep\nDocument results.',
|
||||
]),
|
||||
);
|
||||
await writeFile(
|
||||
join(agentsDir, 'gsd-phase-researcher.md'),
|
||||
makeAgentDef('gsd-phase-researcher', 'Read, Bash', 'You are a researcher.\nSTOP and wait for user input.\nBe thorough.'),
|
||||
);
|
||||
|
||||
const factory = makeFactory();
|
||||
const contextFiles: ContextFiles = { state: '# State' };
|
||||
|
||||
const prompt = await factory.buildPrompt(PhaseType.Research, null, contextFiles);
|
||||
|
||||
// Interactive patterns should be stripped by sanitizePrompt()
|
||||
expect(prompt).not.toContain('AskUserQuestion');
|
||||
expect(prompt).not.toContain('/gsd:');
|
||||
expect(prompt).not.toMatch(/\bSTOP\s+and\s+wait/);
|
||||
|
||||
// Non-interactive content on separate lines should remain
|
||||
expect(prompt).toContain('You are a researcher.');
|
||||
expect(prompt).toContain('Be thorough.');
|
||||
expect(prompt).toContain('Gather data from the project.');
|
||||
expect(prompt).toContain('Analyze findings.');
|
||||
});
|
||||
|
||||
it('buildPrompt with execute+plan sanitizes output from buildExecutorPrompt', async () => {
|
||||
await writeFile(
|
||||
join(agentsDir, 'gsd-executor.md'),
|
||||
makeAgentDef('gsd-executor', 'Read, Write, Edit, Bash', 'You are an executor.\nSTOP and wait for user.\nExecute thoroughly.'),
|
||||
);
|
||||
|
||||
const factory = makeFactory();
|
||||
const plan = makeParsedPlan({ objective: 'Build the auth system' });
|
||||
const contextFiles: ContextFiles = { state: '# State' };
|
||||
|
||||
const prompt = await factory.buildPrompt(PhaseType.Execute, plan, contextFiles);
|
||||
|
||||
// Objective should remain (no interactive pattern on that line)
|
||||
expect(prompt).toContain('Build the auth system');
|
||||
// The role's STOP directive should be stripped
|
||||
expect(prompt).not.toMatch(/\bSTOP\s+and\s+wait/);
|
||||
// Non-interactive role content should remain
|
||||
expect(prompt).toContain('You are an executor.');
|
||||
});
|
||||
});
|
||||
});
|
||||
|
||||
describe('PHASE_WORKFLOW_MAP', () => {
|
||||
|
||||
@@ -8,12 +8,14 @@
|
||||
|
||||
import { readFile } from 'node:fs/promises';
|
||||
import { join } from 'node:path';
|
||||
import { fileURLToPath } from 'node:url';
|
||||
import { homedir } from 'node:os';
|
||||
|
||||
import type { ContextFiles, ParsedPlan } from './types.js';
|
||||
import { PhaseType } from './types.js';
|
||||
import { buildExecutorPrompt, parseAgentRole } from './prompt-builder.js';
|
||||
import { PHASE_AGENT_MAP } from './tool-scoping.js';
|
||||
import { sanitizePrompt } from './prompt-sanitizer.js';
|
||||
|
||||
// ─── Workflow file mapping ───────────────────────────────────────────────────
|
||||
|
||||
@@ -65,16 +67,22 @@ export class PromptFactory {
|
||||
private readonly workflowsDir: string;
|
||||
private readonly agentsDir: string;
|
||||
private readonly projectAgentsDir?: string;
|
||||
private readonly sdkPromptsDir: string;
|
||||
|
||||
constructor(options?: {
|
||||
gsdInstallDir?: string;
|
||||
agentsDir?: string;
|
||||
projectAgentsDir?: string;
|
||||
sdkPromptsDir?: string;
|
||||
}) {
|
||||
const gsdInstallDir = options?.gsdInstallDir ?? join(homedir(), '.claude', 'get-shit-done');
|
||||
this.workflowsDir = join(gsdInstallDir, 'workflows');
|
||||
this.agentsDir = options?.agentsDir ?? join(homedir(), '.claude', 'agents');
|
||||
this.projectAgentsDir = options?.projectAgentsDir;
|
||||
// SDK prompts dir: explicit override → package-relative default via import.meta.url
|
||||
this.sdkPromptsDir =
|
||||
options?.sdkPromptsDir ??
|
||||
join(fileURLToPath(new URL('.', import.meta.url)), '..', 'prompts');
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -91,7 +99,7 @@ export class PromptFactory {
|
||||
// Execute phase with a plan: delegate to existing buildExecutorPrompt
|
||||
if (phaseType === PhaseType.Execute && plan) {
|
||||
const agentDef = await this.loadAgentDef(phaseType);
|
||||
return buildExecutorPrompt(plan, agentDef);
|
||||
return sanitizePrompt(buildExecutorPrompt(plan, agentDef));
|
||||
}
|
||||
|
||||
const sections: string[] = [];
|
||||
@@ -135,17 +143,28 @@ export class PromptFactory {
|
||||
sections.push(`## Phase Instructions\n\n${phaseInstructions}`);
|
||||
}
|
||||
|
||||
return sections.join('\n\n');
|
||||
return sanitizePrompt(sections.join('\n\n'));
|
||||
}
|
||||
|
||||
/**
|
||||
* Load the workflow file for a phase type.
|
||||
* Tries sdk/prompts/workflows/ first (headless versions), then
|
||||
* falls back to GSD-1 originals in workflowsDir.
|
||||
* Returns the raw content, or undefined if not found.
|
||||
*/
|
||||
async loadWorkflowFile(phaseType: PhaseType): Promise<string | undefined> {
|
||||
const filename = PHASE_WORKFLOW_MAP[phaseType];
|
||||
const filePath = join(this.workflowsDir, filename);
|
||||
|
||||
// Try SDK prompts dir first (headless versions)
|
||||
const sdkPath = join(this.sdkPromptsDir, 'workflows', filename);
|
||||
try {
|
||||
return await readFile(sdkPath, 'utf-8');
|
||||
} catch {
|
||||
// Not in sdk/prompts/, fall through to GSD-1 originals
|
||||
}
|
||||
|
||||
// Fall back to GSD-1 originals
|
||||
const filePath = join(this.workflowsDir, filename);
|
||||
try {
|
||||
return await readFile(filePath, 'utf-8');
|
||||
} catch {
|
||||
@@ -155,15 +174,19 @@ export class PromptFactory {
|
||||
|
||||
/**
|
||||
* Load the agent definition for a phase type.
|
||||
* Tries user-level agents dir first, then project-level.
|
||||
* Tries sdk/prompts/agents/ first (headless versions), then
|
||||
* user-level agents dir, then project-level.
|
||||
* Returns undefined if no agent is mapped or file not found.
|
||||
*/
|
||||
async loadAgentDef(phaseType: PhaseType): Promise<string | undefined> {
|
||||
const agentFilename = PHASE_AGENT_MAP[phaseType];
|
||||
if (!agentFilename) return undefined;
|
||||
|
||||
// Try user-level agents dir first
|
||||
const paths = [join(this.agentsDir, agentFilename)];
|
||||
// Try SDK prompts dir first (headless versions)
|
||||
const paths = [
|
||||
join(this.sdkPromptsDir, 'agents', agentFilename),
|
||||
join(this.agentsDir, agentFilename),
|
||||
];
|
||||
|
||||
// Then project-level if configured
|
||||
if (this.projectAgentsDir) {
|
||||
|
||||
@@ -352,7 +352,8 @@ describe('GSDTools typed methods', () => {
|
||||
expect(result.received_args).toContain('init');
|
||||
expect(result.received_args).toContain('phase-op');
|
||||
expect(result.received_args).toContain('7');
|
||||
expect(result.received_args).toContain('--raw');
|
||||
// exec() no longer appends --raw (only execRaw does)
|
||||
expect(result.received_args).not.toContain('--raw');
|
||||
});
|
||||
});
|
||||
|
||||
|
||||
242
sdk/src/prompt-sanitizer.test.ts
Normal file
242
sdk/src/prompt-sanitizer.test.ts
Normal file
@@ -0,0 +1,242 @@
|
||||
import { describe, it, expect } from 'vitest';
|
||||
import { sanitizePrompt } from './prompt-sanitizer.js';
|
||||
|
||||
describe('sanitizePrompt', () => {
|
||||
// ─── Edge cases ──────────────────────────────────────────────────────────
|
||||
|
||||
describe('edge cases', () => {
|
||||
it('returns empty string unchanged', () => {
|
||||
expect(sanitizePrompt('')).toBe('');
|
||||
});
|
||||
|
||||
it('returns undefined/null-ish input unchanged', () => {
|
||||
// eslint-disable-next-line @typescript-eslint/no-explicit-any
|
||||
expect(sanitizePrompt(undefined as any)).toBeUndefined();
|
||||
// eslint-disable-next-line @typescript-eslint/no-explicit-any
|
||||
expect(sanitizePrompt(null as any)).toBeNull();
|
||||
});
|
||||
|
||||
it('preserves clean content with no patterns', () => {
|
||||
const clean = 'This is a clean prompt.\nIt has no interactive patterns.\n\nJust normal text.';
|
||||
expect(sanitizePrompt(clean)).toBe(clean.trim());
|
||||
});
|
||||
});
|
||||
|
||||
// ─── @file: references ───────────────────────────────────────────────────
|
||||
|
||||
describe('@file: references', () => {
|
||||
it('strips lines containing @file: references', () => {
|
||||
const input = 'Before\nLoad @file:path/to/context.md for context\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('@file:');
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('strips @file: with various path formats', () => {
|
||||
const input = [
|
||||
'@file:simple.md',
|
||||
'@file:./relative/path.md',
|
||||
'@file:/absolute/path/to/file.md',
|
||||
'@file:~/.claude/get-shit-done/workflows/execute-plan.md',
|
||||
].join('\n');
|
||||
expect(sanitizePrompt(input)).toBe('');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── /gsd: slash commands ────────────────────────────────────────────────
|
||||
|
||||
describe('/gsd: slash commands', () => {
|
||||
it('strips lines containing /gsd: commands', () => {
|
||||
const input = 'Before\nRun /gsd:execute-plan to proceed\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('/gsd:');
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('strips various /gsd: command formats', () => {
|
||||
const input = [
|
||||
'Use /gsd:research-phase',
|
||||
'Then /gsd:plan-phase --auto',
|
||||
'Finally /gsd:verify-phase',
|
||||
].join('\n');
|
||||
expect(sanitizePrompt(input)).toBe('');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── AskUserQuestion() calls ─────────────────────────────────────────────
|
||||
|
||||
describe('AskUserQuestion() calls', () => {
|
||||
it('strips AskUserQuestion lines', () => {
|
||||
const input = 'Before\nAskUserQuestion("What should we do?")\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('AskUserQuestion');
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('strips AskUserQuestion with various argument styles', () => {
|
||||
const input = [
|
||||
'AskUserQuestion("simple")',
|
||||
'AskUserQuestion( "with spaces" )',
|
||||
' AskUserQuestion("indented")',
|
||||
'Use AskUserQuestion("inline") here',
|
||||
].join('\n');
|
||||
expect(sanitizePrompt(input)).toBe('');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── SlashCommand() calls ────────────────────────────────────────────────
|
||||
|
||||
describe('SlashCommand() calls', () => {
|
||||
it('strips SlashCommand lines', () => {
|
||||
const input = 'Before\nSlashCommand("/gsd:execute")\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('SlashCommand');
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('strips SlashCommand with various forms', () => {
|
||||
const input = [
|
||||
'SlashCommand("proceed")',
|
||||
'SlashCommand( "next" )',
|
||||
' SlashCommand("indented")',
|
||||
].join('\n');
|
||||
expect(sanitizePrompt(input)).toBe('');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── STOP directives ────────────────────────────────────────────────────
|
||||
|
||||
describe('STOP directives', () => {
|
||||
it('strips "STOP and wait" lines', () => {
|
||||
const input = 'Before\nSTOP and wait for user input\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('STOP');
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('strips bare STOP lines', () => {
|
||||
const input = 'Before\nSTOP\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('STOP');
|
||||
});
|
||||
|
||||
it('strips STOP with trailing punctuation', () => {
|
||||
const input = 'Before\nSTOP.\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('STOP');
|
||||
});
|
||||
|
||||
it('strips "STOP here" and "STOP now"', () => {
|
||||
const input = 'STOP here\nSTOP now\nSTOP and ask';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).toBe('');
|
||||
});
|
||||
|
||||
it('preserves STOP in normal prose (not as directive)', () => {
|
||||
const input = 'Do not stop the build process.';
|
||||
const result = sanitizePrompt(input);
|
||||
// "stop" in lowercase in normal prose should be preserved
|
||||
expect(result).toContain('stop the build');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── 'wait for user' / 'ask the user' instructions ──────────────────────
|
||||
|
||||
describe('wait for user / ask the user', () => {
|
||||
it('strips "wait for user" lines', () => {
|
||||
const input = 'Before\nWait for user confirmation before proceeding\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toMatch(/wait for.*user/i);
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('strips "wait for the user" lines', () => {
|
||||
const input = 'Before\nWait for the user to respond\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toMatch(/wait for the user/i);
|
||||
});
|
||||
|
||||
it('strips "ask the user" lines', () => {
|
||||
const input = 'Before\nAsk the user for clarification\nAfter';
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toMatch(/ask the user/i);
|
||||
expect(result).toContain('Before');
|
||||
expect(result).toContain('After');
|
||||
});
|
||||
|
||||
it('is case-insensitive for wait/ask patterns', () => {
|
||||
const input = [
|
||||
'WAIT FOR USER input',
|
||||
'wait for user approval',
|
||||
'ASK THE USER what to do',
|
||||
'ask the user for feedback',
|
||||
].join('\n');
|
||||
expect(sanitizePrompt(input)).toBe('');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── Multiple patterns in one string ─────────────────────────────────────
|
||||
|
||||
describe('multiple patterns in one string', () => {
|
||||
it('strips all pattern types from a mixed prompt', () => {
|
||||
const input = [
|
||||
'## Research Phase',
|
||||
'',
|
||||
'Investigate the codebase using @file:context.md for context.',
|
||||
'',
|
||||
'When done, run /gsd:plan-phase to proceed.',
|
||||
'',
|
||||
'If unclear, AskUserQuestion("What should I focus on?")',
|
||||
'',
|
||||
'STOP and wait for user input.',
|
||||
'',
|
||||
'Use SlashCommand("next") to continue.',
|
||||
'',
|
||||
'Wait for user confirmation before executing.',
|
||||
'',
|
||||
'This line is clean and should remain.',
|
||||
].join('\n');
|
||||
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).not.toContain('@file:');
|
||||
expect(result).not.toContain('/gsd:');
|
||||
expect(result).not.toContain('AskUserQuestion');
|
||||
expect(result).not.toContain('SlashCommand');
|
||||
expect(result).not.toMatch(/\bSTOP\b/);
|
||||
expect(result).not.toMatch(/wait for user/i);
|
||||
expect(result).toContain('## Research Phase');
|
||||
expect(result).toContain('This line is clean and should remain.');
|
||||
});
|
||||
});
|
||||
|
||||
// ─── Blank line collapsing ───────────────────────────────────────────────
|
||||
|
||||
describe('blank line collapsing', () => {
|
||||
it('collapses 3+ consecutive blank lines to 2', () => {
|
||||
const input = 'Line 1\n\n\n\n\nLine 2';
|
||||
const result = sanitizePrompt(input);
|
||||
// After trim(), the result should have at most 2 consecutive newlines
|
||||
expect(result).not.toMatch(/\n{3,}/);
|
||||
expect(result).toContain('Line 1');
|
||||
expect(result).toContain('Line 2');
|
||||
});
|
||||
|
||||
it('collapses blanks left by stripped lines', () => {
|
||||
const input = [
|
||||
'Before',
|
||||
'',
|
||||
'AskUserQuestion("something")',
|
||||
'',
|
||||
'After',
|
||||
].join('\n');
|
||||
const result = sanitizePrompt(input);
|
||||
expect(result).toBe('Before\n\nAfter');
|
||||
});
|
||||
});
|
||||
});
|
||||
71
sdk/src/prompt-sanitizer.ts
Normal file
71
sdk/src/prompt-sanitizer.ts
Normal file
@@ -0,0 +1,71 @@
|
||||
/**
|
||||
* Prompt sanitizer — strips interactive CLI patterns from GSD-1 prompts
|
||||
* so they're safe for headless SDK use.
|
||||
*
|
||||
* Patterns removed:
|
||||
* - @file:... references (file injection directives)
|
||||
* - /gsd:... slash commands
|
||||
* - AskUserQuestion(...) calls
|
||||
* - STOP directives in interactive contexts
|
||||
* - SlashCommand() calls
|
||||
* - 'wait for user' / 'ask the user' instructions
|
||||
*/
|
||||
|
||||
// ─── Pattern definitions ─────────────────────────────────────────────────────
|
||||
|
||||
/**
|
||||
* Each pattern is a regex that matches a full line (or inline span) to remove.
|
||||
* We strip matching lines entirely to avoid leaving blank gaps that break
|
||||
* markdown structure.
|
||||
*/
|
||||
const LINE_PATTERNS: RegExp[] = [
|
||||
// @file:path/to/something references — entire line
|
||||
/^.*@file:\S+.*$/gm,
|
||||
|
||||
// /gsd:command references — entire line containing a slash command
|
||||
/^.*\/gsd:\S+.*$/gm,
|
||||
|
||||
// AskUserQuestion(...) calls — entire line
|
||||
/^.*AskUserQuestion\s*\(.*$/gm,
|
||||
|
||||
// SlashCommand() calls — entire line
|
||||
/^.*SlashCommand\s*\(.*$/gm,
|
||||
|
||||
// STOP directives — lines that are primarily "STOP" instructions
|
||||
// Match lines where STOP is used as an imperative (not as part of normal prose)
|
||||
/^.*\bSTOP\b(?:\s+(?:and\s+)?(?:wait|ask|here|now)).*$/gm,
|
||||
/^\s*STOP\s*[.!]?\s*$/gm,
|
||||
|
||||
// 'wait for user' / 'ask the user' instruction lines
|
||||
/^.*\bwait\s+for\s+(?:the\s+)?user\b.*$/gim,
|
||||
/^.*\bask\s+the\s+user\b.*$/gim,
|
||||
];
|
||||
|
||||
// ─── Public API ──────────────────────────────────────────────────────────────
|
||||
|
||||
/**
|
||||
* Strip interactive CLI patterns from a prompt string.
|
||||
*
|
||||
* Removes lines matching known interactive patterns (file references,
|
||||
* slash commands, user-interaction directives) while preserving all
|
||||
* other content unchanged.
|
||||
*
|
||||
* @param input - Raw prompt string, possibly containing interactive patterns
|
||||
* @returns Cleaned prompt with interactive patterns removed
|
||||
*/
|
||||
export function sanitizePrompt(input: string): string {
|
||||
if (!input) return input;
|
||||
|
||||
let result = input;
|
||||
|
||||
for (const pattern of LINE_PATTERNS) {
|
||||
// Reset lastIndex for global regexes
|
||||
pattern.lastIndex = 0;
|
||||
result = result.replace(pattern, '');
|
||||
}
|
||||
|
||||
// Collapse runs of 3+ blank lines down to 2 (preserve paragraph breaks)
|
||||
result = result.replace(/\n{3,}/g, '\n\n');
|
||||
|
||||
return result.trim();
|
||||
}
|
||||
Reference in New Issue
Block a user