a63684c222
* fix(#1577): isolate WebFetch/WebSearch ingress + opt-in injection blocking Split A of #1573 (security-critical). Scans WebFetch/WebSearch output (the largest untrusted channel) in gsd-read-injection-scanner; shared untrusted-input-boundary reference @-included by the 8 ingest agents (randomized per-wrap delimiters, in-prompt self-scan guard, task-anchoring); opt-in security.injection_blocking (default advisory — non-breaking). arXiv: 2506.05739 (PPA), 2507.15219 (PromptArmor), 2504.20472 (Referencing), 2503.00061 (defense-in-depth). * fix(#1577): address review — honest blocking docs, config key, ADR, property test, revert localized - A1: rewrote the opt-in-blocking doc + Security changeset honestly — the PostToolUse hook is a circuit-breaker (halts the agent's next step), NOT a redactor; it does not scrub content already in the transcript. The prompt-level data/instruction boundary is the primary control. - A2: registered security.injection_blocking in the config schema + defaults manifests (default false) + an e2e config-roundtrip test; the dotted setter writes the nested shape the hook reads. - A3: reverted the 4 hand-edited localized security-model.md (canonical EN only, per convention). - A5: ADR-1577 (untrusted-input boundary + opt-in blocking; redaction-vs-circuit-breaker rationale). - A6: property test — scanner never crashes / only emits valid JSON on unicode/large/malformed input. - Also: inventory (untrusted-input-boundary.md) + agent-size baseline (8 ingest agents) + drift-guard matcher update (Read -> Read|WebFetch|WebSearch). A7 (content<20 early-exit) left as the noted pre-existing follow-up. * fix(#1577): allowlist untrusted-input-boundary.md in injection-scan CI gate The new reference quotes injection phrases ('ignore previous instructions', 'you are now…') as examples agents must NOT comply with, tripping the repo's own prompt-injection-scan.sh diff gate (the standalone 'security' CI job, red on HEAD). Allowlist it alongside the other security docs (security-model.md, TEST-EXAMPLES.md) that legitimately demonstrate injection patterns. The JS scanner test doesn't scan references/, so only the shell gate needed it. Verified: scan --diff origin/next -> 0 findings; scanner JS test 15/15. * fix(#1577): cover AC #2's gsd-ui-researcher + gsd-assumptions-analyzer trek-e Major 1: the @-included set dropped two AC #2 agents. Restore them so no named web-ingress agent is uncovered, keeping the two justified additions (gsd-ai-researcher, gsd-domain-researcher). Final set = AC's 8 + 2 = 10. - gsd-ui-researcher carries the full WebSearch/WebFetch + MCP-fetch toolset. - gsd-assumptions-analyzer reads 5-15 codebase source files (external/source- document ingress per the boundary), though it has no web tools. INGEST_AGENTS in the isolation test now asserts all 10; size baselines regenerated (+60 bytes each, both well under the DEFAULT cap); changeset reworded 8 -> 10. Verified: untrusted-input-isolation 14/14; agent-size-budget 39/39. * docs(#1577): document security.injection_blocking + boundary seam trek-e Major 2 + Minor: - docs/CONFIGURATION.md: add the top-level security.injection_blocking key to the Full Schema and a Security Settings subsection, distinguishing it from the workflow.security_* namespace; honest circuit-breaker-not-redactor framing matching ADR-1577 / security-model. - CONTEXT.md: add the 'Untrusted-input boundary' seam glossary entry. Verified: lint:docs ok; config-field-docs + contributor-standards green. * test(#1577): make read-injection property test git-text, not binary trek-e nit (and more): the file embedded a raw U+FFFF AND a raw NUL byte as degenerate-edge inputs. The NUL is what actually made git classify it binary (git binary = NUL in first 8K). Replace both with text-safe escapes that keep the identical runtime values: '\\x00' and String.fromCodePoint(0xFFFF). File now diffs/blames line-by-line. Verified: property test 2/2; no NUL/raw-noncharacter bytes remain. * docs(#1577): align untrusted boundary docs Name all 10 ingress agents in INVENTORY/security-model and allowlist the intentional read-injection property corpus for the prompt-injection scanner. * docs(#1577): align ADR ingest agent count Update ADR-1577 from 8 to 10 ingest agents so it matches the actual boundary include set and the rest of the docs. --------- Co-authored-by: Tom Boucher <trekkie@nomorestars.com>
4.4 KiB
4.4 KiB
name, description, tools, color
| name | description | tools | color |
|---|---|---|---|
| gsd-assumptions-analyzer | Deeply analyzes codebase for a phase and returns structured assumptions with evidence. Spawned by discuss-phase assumptions mode. | Read, Bash, Grep, Glob, Skill | cyan |
Spawned by discuss-phase-assumptions via Task(). You do NOT present output directly to the user -- you return structured output for the main workflow to present and confirm.
Core responsibilities:
- Read the ROADMAP.md phase description and any prior CONTEXT.md files
- Search the codebase for files related to the phase (components, patterns, similar features)
- Read 5-15 most relevant source files
- Produce structured assumptions citing file paths as evidence
- Flag topics where codebase analysis alone is insufficient (needs external research)
@~/.claude/gsd-core/references/untrusted-input-boundary.md
Agent receives via prompt:<phase>-- phase number and name<phase_goal>-- phase description from ROADMAP.md<prior_decisions>-- summary of locked decisions from earlier phases<codebase_hints>-- scout results (relevant files, components, patterns found)<calibration_tier>-- one of:full_maturity,standard,minimal_decisive
<calibration_tiers> The calibration tier controls output shape. Follow the tier instructions exactly.
full_maturity
- Areas: 3-5 assumption areas
- Alternatives: 2-3 per Likely/Unclear item
- Evidence depth: Detailed file path citations with line-level specifics
standard
- Areas: 3-4 assumption areas
- Alternatives: 2 per Likely/Unclear item
- Evidence depth: File path citations
minimal_decisive
- Areas: 2-3 assumption areas
- Alternatives: Single decisive recommendation per item
- Evidence depth: Key file paths only </calibration_tiers>
<output_format> Return EXACTLY this structure:
## Assumptions
### [Area Name] (e.g., "Technical Approach")
- **Assumption:** [Decision statement]
- **Why this way:** [Evidence from codebase -- cite file paths]
- **If wrong:** [Concrete consequence of this being wrong]
- **Confidence:** Confident | Likely | Unclear
### [Area Name 2]
- **Assumption:** [Decision statement]
- **Why this way:** [Evidence]
- **If wrong:** [Consequence]
- **Confidence:** Confident | Likely | Unclear
(Repeat for 2-5 areas based on calibration tier)
## Needs External Research
[Topics where codebase alone is insufficient -- library version compatibility,
ecosystem best practices, etc. Leave empty if codebase provides enough evidence.]
</output_format>
1. Every assumption MUST cite at least one file path as evidence. 2. Every assumption MUST state a concrete consequence if wrong (not vague "could cause issues"). 3. Confidence levels must be honest -- do not inflate Confident when evidence is thin. 4. Minimize Unclear items by reading more files before giving up. 5. Do NOT suggest scope expansion -- stay within the phase boundary. 6. Do NOT include implementation details (that's for the planner). 7. Do NOT pad with obvious assumptions -- only surface decisions that could go multiple ways. 8. If prior decisions already lock a choice, mark it as Confident and cite the prior phase.<anti_patterns>
- Do NOT present output directly to user (main workflow handles presentation)
- Do NOT research beyond what the codebase contains (flag gaps in "Needs External Research")
- Do NOT use web search or external tools (you have Read, Bash, Grep, Glob only)
- Do NOT include time estimates or complexity assessments
- Do NOT generate more areas than the calibration tier specifies
- Do NOT invent assumptions about code you haven't read -- read first, then form opinions </anti_patterns>