Files
msd-core/tests/debug-manager-blocking-spawn.test.cjs
Tom Boucher 6f99e493e7 fix(#4395): make the debug session manager's own gsd-debugger spawn blocking (#4718)
* test(#4395): prove the manager spawns its debugger without blocking

Failing-first regression coverage for #4395.

debug.md:209 mandates the orchestrator to session-manager spawn carry
run_in_background: false, and says why outright: "Claude Code backgrounds
subagents by default, and only that flag makes the spawn return the
compact session summary directly" (#2196).

The session-manager to debugger spawn, one level down, carries no flag.
Measured: run_in_background appears nowhere under agents/ -- only in
gsd-core/workflows/. So by the rule #2196 itself states, that spawn is
backgrounded, Step 3 ("Handle Agent Return") has no return to inspect,
the manager emits CONTINUE_REQUIRED, the orchestrator auto-resumes per
#2257/#3448, and a second detached debugger races the first on
.planning/debug/<slug>.md.

Row 4 is the load-bearing one: it closes the CLASS by requiring every
subagent spawn under agents/ to declare run_in_background explicitly, so
the next agent that spawns one has to decide rather than inherit a silent
host default. It is scoped to agents/ precisely so it cannot misfire on
the workflows that deliberately use true for parallel fan-out.

Rows 5-7 are pins, not fixes: the #2196 mandate one level up, Step 2 as
the single spawn-format source that the eight continuation sites delegate
to, and the survival of CONTINUE_REQUIRED (which has a legitimate trigger
unrelated to this defect).

Red round: 4 of 7 rows fail. Row 3 needed hardening first -- asserting
only that the two variants AGREE passed vacuously, because two missing
flags are also equal; it now asserts each is present before comparing.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix(#4395): make the manager's own debugger spawn blocking

The orchestrator-to-manager hop already requires a blocking spawn and says
why (#2196, debug.md:209): Claude Code backgrounds subagents by default,
and only run_in_background=false makes the spawn return its summary. The
manager-to-debugger hop, one level down, carried no flag -- measured,
run_in_background appeared nowhere under agents/ at all.

So that spawn was backgrounded. Step 3 ("Handle Agent Return") opens
"Inspect the return output for the structured return header" -- with
nothing to inspect, the manager correctly declined to fabricate a terminal
summary and returned CONTINUE_REQUIRED; the orchestrator correctly
auto-resumed (#2257/#3448); the resumed manager reached Step 2 and spawned
a SECOND detached debugger. Both then raced on .planning/debug/<slug>.md.

Every observable in the report follows with no further assumption,
including the count: the reporter saw exactly three collisions in one
invocation, and debug.md:251 caps auto-resumes at three per slug -- one
collision per cycle.

Fixed at the cause, in both shipped variants, kept byte-consistent. The
eight continuation sites say "see Step 2 format", so they inherit it.

The issue offered two remedies. The second -- have the auto-resume path
reconcile a still-running debugger before spawning another -- is not taken:
it treats the symptom, and needs machinery that does not exist (no portable
way to enumerate or stop another runtime's live agents, plus an in-flight
sentinel with staleness and recovery rules, or an orphaned marker deadlocks
the session permanently). With the spawn blocking, the manager cannot reach
Step 4 while a debugger is live, so such a guard would also be unreachable.

#2257, #3448, the anti-loop heuristic, the cap of three, and the
CONTINUE_REQUIRED shape are all correct and untouched. CONTINUE_REQUIRED
keeps its legitimate trigger: the manager genuinely exhausting its own turn
budget mid-investigation.

Also corrects the red-round test to the canonical CALL form. debug.md
writes run_in_background=false inside Agent(...) and run_in_background:
false in prose; the first draft asserted the prose form, which the shipped
call would never have matched.

Emitted-Drift-Ack-Growth: gsd-debug-session-manager.md — the blocking spawn flag plus the note recording why an unstated flag produced colliding debuggers
Emitted-Drift-Ack-Growth: gsd-debug-session-manager.compact.md — same change as its full sibling, kept byte-consistent with it

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* chore(#4395): add changeset fragment

pr:0 placeholder is backfilled with the real number once the PR exists.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* chore(#4395): refresh the variant benchmark baseline

Two entries move.

gsd-debug-session-manager.md 4766/4477 -> 4938/4649 is this change: the
blocking-spawn flag plus its explanatory note, added to BOTH variants to
keep them byte-consistent, so the compact sibling grows by the same amount
and the pair's reduction ratio dips 6.06 -> 5.85. The compact file remains
strictly smaller than its canonical sibling, which is what the variant
guard's size check actually requires.

gsd-code-fixer.md 10741 -> 10740 is NOT from this branch -- the file is
untouched here. It has scored 10740 since f334f277dd (#4324) reworded a
line without refreshing this fixture, so the stale number is sitting on
next. Fixed here rather than deferred; the #4350 branch carries the
identical one-token correction, so whichever lands first makes the other a
no-op.

Refreshed with the variant script's own --write.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix(#4395): state the blocking rule agent-wide and correct two overclaims

Review round. Three substantive corrections, one of them to a claim I made
in the previous commit message.

1. The blocking rule is now stated AGENT-WIDE, not per-call. My earlier
   claim that "the eight continuation sites say 'see Step 2 format', so
   they inherit it" was false: exactly ONE of them names Step 2 (the
   compact variant says "Step 2 format" without the "see", which is why
   the first draft of the test matched it zero times there). The other
   sites inherit only because Step 2 holds the sole Agent() spawn literal
   in each file. Both variants now say so outright, and the test pins the
   sole-literal invariant in BOTH variants rather than the prose wording
   in one.

2. debug.md's two auto-resume buckets now restate run_in_background=false
   for the re-spawn. "The same session_params" does not carry it --
   session_params is prompt content, not the spawn flag.

3. The test extractor now also matches single-line Agent(...) calls. The
   class guard was blind to exactly the shape a future offender is most
   likely to take.

Also corrects the diagnosis: remedy 2 is NOT unreachable once the spawn
blocks. This agent's own retained CONTINUE_REQUIRED trigger is "turn
budget exhausted WHILE THE DEBUGGER IS STILL INVESTIGATING", so a harness
turn cutoff mid-wait still double-spawns; debug.md:209 names the same
class from the other side. What this fix removes is the SYSTEMATIC case --
every invocation, because the spawn was always backgrounded. The residual
turn-cutoff window survives, bounded by the existing three-resume cap, and
is stated in the PR rather than denied.

Emitted-Drift-Ack-Growth: gsd-debug-session-manager.md — blocking spawn flag, the note recording why an unstated flag produced colliding debuggers, and the agent-wide restatement the per-site inheritance actually depends on
Emitted-Drift-Ack-Growth: gsd-debug-session-manager.compact.md — same change as its full sibling, kept byte-consistent with it
Emitted-Drift-Ack-Growth: debug.md — both auto-resume buckets restate the spawn flag, since session_params does not carry it
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* chore(#4395): backfill the changeset PR number

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: sim <sim@local>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-09-14 02:12:46 -04:00

147 lines
6.9 KiB
JavaScript

// Shipped agent Markdown is the installed contract: the runtime reads these
// files and performs the Agent() spawns they describe, so the shipped text is
// the subject under test.
'use strict';
const { test, describe } = require('node:test');
const assert = require('node:assert/strict');
const fs = require('node:fs');
const path = require('node:path');
const { readFileNormalized } = require('./helpers.cjs');
const ROOT = path.join(__dirname, '..');
const AGENTS_DIR = path.join(ROOT, 'agents');
const MANAGER = path.join(AGENTS_DIR, 'gsd-debug-session-manager.md');
const MANAGER_COMPACT = path.join(AGENTS_DIR, 'gsd-debug-session-manager.compact.md');
const DEBUG_WORKFLOW = path.join(ROOT, 'gsd-core', 'workflows', 'debug.md');
// Pull out each `Agent( … )` call literally, so a row asserts against the
// shipped spawn rather than against a pattern retyped from the fix.
function agentSpawnBlocks(content) {
const blocks = [];
// Multi-line form: Agent(\n … \n). Non-greedy, so it stops at the first
// line that is exactly ")".
const multi = /Agent\(\s*\n[\s\S]*?\n\)/g;
// Single-line form: Agent(…). Without this a one-line spawn evades the
// class guard in row 4 entirely — the guard would be vacuous against
// exactly the kind of future offender it exists to catch.
const single = /Agent\([^\n)]*\)/g;
let m;
while ((m = multi.exec(content)) !== null) blocks.push(m[0]);
while ((m = single.exec(content)) !== null) blocks.push(m[0]);
return blocks;
}
function debuggerSpawn(content, label) {
const matches = agentSpawnBlocks(content).filter((b) => b.includes('subagent_type="gsd-debugger"'));
assert.equal(matches.length, 1, `${label}: expected exactly one gsd-debugger spawn block`);
return matches[0];
}
describe('debug session manager: the debugger spawn must block (#4395)', () => {
// ROW 1 + 2 — the reported defect, in both shipped variants.
for (const [label, file] of [['full', MANAGER], ['compact', MANAGER_COMPACT]]) {
test(`${label} variant spawns gsd-debugger with run_in_background: false`, () => {
const spawn = debuggerSpawn(readFileNormalized(file), `${label} variant`);
// Claude Code backgrounds subagents by DEFAULT — debug.md:209 (#2196) says
// so outright. Without the flag the manager's Step 3 ("Handle Agent
// Return") has no return to inspect, so it emits CONTINUE_REQUIRED, the
// orchestrator auto-resumes, and a second detached debugger races the
// first on .planning/debug/<slug>.md.
assert.match(
spawn,
/run_in_background\s*=\s*false/,
`${label} variant: the debugger spawn must be blocking, or the manager cannot see its result`,
);
});
}
// ROW 3 — the pair cannot drift apart.
test('both variants agree on the debugger spawn flag', () => {
const full = debuggerSpawn(readFileNormalized(MANAGER), 'full variant');
const compact = debuggerSpawn(readFileNormalized(MANAGER_COMPACT), 'compact variant');
const flagOf = (s) => (s.match(/run_in_background\s*=\s*(true|false)/) || [])[1];
// Assert the flag is PRESENT before asserting agreement: two missing flags
// are also "equal", so a bare equality check passes vacuously on the very
// tree this test exists to reject.
assert.ok(flagOf(full), 'full variant must declare the flag at all');
assert.ok(flagOf(compact), 'compact variant must declare the flag at all');
assert.equal(
flagOf(compact),
flagOf(full),
'the compact sibling must not diverge from the canonical agent on spawn semantics',
);
});
// ROW 4 — close the CLASS, not just this instance.
test('every subagent spawn under agents/ declares run_in_background explicitly', () => {
const offenders = [];
for (const name of fs.readdirSync(AGENTS_DIR)) {
if (!name.endsWith('.md')) continue;
const file = path.join(AGENTS_DIR, name);
for (const block of agentSpawnBlocks(readFileNormalized(file))) {
if (!block.includes('subagent_type=')) continue;
if (!/run_in_background\s*=\s*(true|false)/.test(block)) {
offenders.push(`${name}: ${block.split('\n').find((l) => l.includes('subagent_type=')).trim()}`);
}
}
}
assert.deepEqual(
offenders,
[],
'An agent that spawns a subagent must DECIDE whether that spawn blocks. ' +
'Leaving it unstated inherits the host default (Claude Code backgrounds ' +
'subagents), which is how #4395 produced colliding debuggers.\n ' +
offenders.join('\n '),
);
});
// ROW 5 — the fix one level up is undisturbed.
test('the orchestrator to session-manager hop still mandates a blocking spawn (#2196)', () => {
const workflow = readFileNormalized(DEBUG_WORKFLOW);
assert.match(
workflow,
/run_in_background:\s*false`\s*—\s*Claude Code backgrounds subagents by default/,
'debug.md must keep the #2196 mandate that the session-manager spawn blocks',
);
});
// ROW 6 — Step 2 must stay the SOLE spawn literal, in both variants.
//
// This is the invariant the fix actually rests on, and it is narrower than it
// first looks. The continuation sites do NOT all name Step 2: in the full
// variant, 8 sites say "spawn continuation agent" and exactly ONE adds
// "(see Step 2 format)"; the compact variant says "Step 2 format" without the
// "see". The other sites inherit the blocking flag only because there is no
// other Agent() spawn literal in the file to inherit from. Pin that, not the
// prose wording — if a second literal is ever added, those sites silently stop
// inheriting and #4395 comes back on whichever paths route through it.
for (const [label, file] of [['full', MANAGER], ['compact', MANAGER_COMPACT]]) {
test(`${label} variant keeps Step 2 as the sole Agent() spawn literal`, () => {
const content = readFileNormalized(file);
const spawnBlocks = agentSpawnBlocks(content).filter((b) => b.includes('subagent_type='));
assert.equal(
spawnBlocks.length,
1,
`${label} variant: exactly one Agent() spawn literal may exist — the continuation sites ` +
're-spawn by reference to it, so a second literal would not carry the flag',
);
assert.ok(
/Step 2 format/.test(content),
`${label} variant: at least one continuation site must still name Step 2 as the format source`,
);
});
}
// ROW 7 — the non-terminal shape was not collaterally removed.
test('CONTINUE_REQUIRED and both terminal markers survive', () => {
const manager = readFileNormalized(MANAGER);
// CONTINUE_REQUIRED has a legitimate trigger unrelated to this defect: the
// manager genuinely exhausting its own turn budget mid-investigation.
assert.match(manager, /## CONTINUE_REQUIRED/, 'the non-terminal marker must remain');
assert.match(manager, /## DEBUG SESSION COMPLETE/, 'the terminal marker must remain');
assert.match(manager, /ABANDONED/, 'the abandoned disposition must remain');
});
});