Files
msd-core/gsd-core/references/failing-direction.md
Tom Boucher c933184b97 enhance(#3172): require a stated failing direction for every automated acceptance command (#3825)
* test(#3172): failing-first suite for the stated failing-direction probe

Pins the <fails_when> pairing walk, placeholder denylist, MISSING sentinel
exemption, degraded-read contract, CLI arm and the plan-authoring contract text.
RED by construction: the module exports it requires do not exist yet.
Executed on the remote runner.

* feat(#3172): require a stated failing direction for every automated acceptance command

Every runnable <automated> command now carries a <fails_when> sibling naming
what output constitutes failure. A command with no expressible failure mode is
not an acceptance test: it reads as rigour and is not falsifiable.

- verify-command-grounding gains a failing-direction probe sharing the existing
  <automated> grammar, MISSING sentinel and walk guard rather than copying them
- gsd-tools check verify-failure-directions <N> backs it; plan-phase dispatches
  it and hands the JSON to gsd-plan-checker check 8f
- Dimension 8 detail extracted to references to stay under the agent size cap

Verified on the remote runner.

* fix(#3172): close four review findings in the failing-direction probe

- MISSING_SENTINEL_RE matched an env-var assignment prefix (MISSING=1 cmd), so
  a real command was exempted from the new blocking gate. Tightened the SHARED
  constant rather than adding a second copy.
- Both token regexes scanned to EOF on unclosed openers (O(n^2), 1562ms at 40k).
  Bodies are now non-crossing; 1ms, byte-identical on well-formed input. The
  pre-existing AUTOMATED_BLOCK_RE carried the same defect and is fixed here too.
- probePhaseFailingDirections reported status 'ok' when one plan was unreadable,
  conflating 'could not look' with 'nothing to report'.
- Extracted the phase-resolution block both check arms had copied verbatim.

Also corrects a docs/AGENTS.md dimension list stale since #2401.
Verified on the remote runner.

* fix(#3172): project the planner rule onto the spawn contract, settle emitted bookkeeping

The remote runner refuted the planner-side edit. agents/gsd-planner.md is frozen
under a 49152-LF-char cap asserted by four suites and sat at 49,146 — six chars
of headroom — so the +537 of authoring rule blew it. #3297/#3645 already settled
where such a rule goes: the planner spawn contract in plan-phase.md, beside
<tracked_source_paths>. The agent file is reverted to origin/next verbatim.

- plan-phase.md gains <failing_direction_contract>; tests row 30 now asserts the
  contract there and row 30b guards the freeze in both directions
- plan-phase.md growth acknowledged by APPENDING to the 3409 fragment, per the
  precedent that two ack sources may never name the same path
- install-tree fixtures regenerated for the three new reference files

Verified on the remote runner.

* chore(#3172): backfill PR number into the changeset fragment

pr:0 -> pr:3825 now that the PR exists.

---------

Co-authored-by: sim <sim@local>
2026-08-24 19:05:11 -04:00

79 lines
4.2 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# Check 8f — Stated Failing Direction (#3172)
> Reference file for gsd-plan-checker agent. Loaded on-demand via `@` reference.
**Question:** For each runnable `<automated>` command, does the plan say what output constitutes
failure?
Checks 8a–8d ask whether an acceptance command is *present*, and Verify Command Path
Resolvability asks whether its target *resolves*. This one asks whether the command is
**falsifiable at all**. A command with no expressible failure mode is not an acceptance test —
it reads as rigour and delivers none.
#3172: a planner emitted 21 `<automated>` commands that could not run. Cargo exited non-zero, so
that instance failed loudly — luck, not design. The same class of error with a command that
exits 0 on a no-op passes green and silently. Requiring a stated failing direction is the only
shape that catches the silent case, because it forces the plan to name the failure signal rather
than assume the command has one.
## The contract
```xml
<verify>
<automated>npm --prefix apps/api test -- auth.spec.ts</automated>
<fails_when>non-zero exit, or "0 passed" in the summary line</fails_when>
</verify>
```
Within one `<task>`, each `<fails_when>` binds to the nearest **preceding** `<automated>`; a
command's binding statement is the **first** one that follows it. N runnable commands need N
statements.
## Do not hand-reason this
`gsd-core/workflows/plan-phase.md` runs the deterministic probe **before** spawning this checker
and interpolates the result into the verification prompt as `{FAILING_DIRECTIONS}`, inside a
`<failing_direction_probe>` block. This check reads that already-supplied JSON — it never invokes
`gsd_run check verify-failure-directions` itself. If `{FAILING_DIRECTIONS}` is absent from the
prompt, treat this check as silent (nothing to check) rather than trying to run the probe.
The probe never executes command text (PLAN.md is untrusted, LLM-authored). It is a **presence**
recognizer: it proves a statement exists and is not a placeholder. It does not judge whether the
statement names the *right* signal — that judgment is yours, below.
## Process — act on `severity` only
| `severity` | `status` | Action |
|---|---|---|
| `blocker` | `missing` | **BLOCKER** — quote the `command` verbatim: it has no stated failing direction |
| `blocker` | `empty` | **BLOCKER** — a `<fails_when>` is present but blank |
| `blocker` | `placeholder` | **BLOCKER** — quote the placeholder text; `TBD` is not a failure signal |
| `warning` | `orphan` | **WARNING** — a `<fails_when>` that follows no command; it satisfies nothing |
| `none` | `ok` / `sentinel` | silent |
Rules:
- **Report, never prescribe.** State which command has no stated failure mode. Do **not** author
the statement for the planner. A prescribed statement is copied verbatim and carries zero
information — that reproduces #3172 one level up, exactly as the #2401 probe refuses to
prescribe a replacement path.
- `status: sentinel` is a `MISSING — Wave 0 must create …` placeholder command. It is not
runnable, so it has no failure mode to state. **Not a finding.** Say nothing; checks 8a/8d
own it.
- A non-empty `readError` means the probe **could not look**. Report that as a WARNING in its own
words; it is not a clean bill of health.
- **Your added judgment, on `ok` rows only:** a statement that is present, non-empty and
non-placeholder can still be vacuous — *"the command fails"*, *"it doesn't work"*, *"an error
occurs"* restate the word "failure" without naming an observable signal. Raise those as a
**WARNING**, naming what a usable statement looks like (an exit code, a string in the output, a
missing line). Do **not** escalate a vacuous statement to BLOCKER: the deterministic layer owns
the blockers so that a BLOCKER is always reproducible, and prose judgment stays advisory.
- **Length is not a signal.** `<fails_when>non-zero exit</fails_when>` is a complete failing
direction. There is no minimum length, word count, or required keyword.
## Not in scope
A command that runs successfully and asserts nothing (a test-name filter matching zero tests and
exiting 0) is the adjacent **vacuous pass** problem. It is explicitly out of scope for this
check — see the issue's "Out of scope". Do not report it here.