| name | ai-challenge |
| description | Attacks a written specification from outside it: executes the sentences it claims are true, and reports where the tree disagrees. Trigger for "challenge this spec", "attack this decision", "is this spec true", "grill this before I approve it". Not for judging a diff — use /ai-review, which reads a change against a spec and cannot also be its accuser. Not for finding a cause in code — use /ai-debug. It never approves, revises or rewrites: it produces findings that carry a command, and a person decides what they mean. |
| license | Apache-2.0 |
| compatibility | needs git |
| context | fork |
| background | false |
| disable-model-invocation | true |
Execute the sentence, do not re-read it
What it produces
specs/NNN-slug/challenge.md, a list of findings. Every finding names a sentence in the
specification, the command that tested it, and what the command actually printed. A finding
with no command is deleted before the file is written: an opinion about a document is what
the document's own author already had.
Why it is not the section inside the spec
ai-spec requires the author to challenge their own recommendation once, and that section is
worth keeping — but the questioner and the answerer are the same reader, and a reader can
only reason about what they already wrote. Measured on this repository: four of twenty specs
carry that section, and not one of the four contains a command or a file reference. Three
claims in one of them were false, and executing them is what found it.
Steps
- Read only the specification and the tree. Not the plan, not the pull request, not the
conversation that produced it — a challenger who reads the author's reasoning inherits it.
- List every sentence that asserts something checkable: a count, a path, a behaviour, a
claim about another document. Those are the targets; the rest is intent and not yours.
- Execute each one. A count gets counted, a path gets opened, a behaviour gets run. Paste
what came back, including when it agrees.
- Write the findings, worst first.
WRONG when the tree says otherwise, UNPROVEN when
nothing in the tree can decide it, and say which — those are different problems for the
author.
The loop is bounded
At most two rounds against the same spec digest — the canonical bytes ai-eng spec show
prints. A revision changes the digest and reopens the count; the second round against an
unchanged digest is the last. At the ceiling, write the outstanding findings worst first
and hand the page to the person. loopgate is the orchestrator's instrument, not yours:
the skill layer's bound is this one, and the two-identical-greens rule is what an
automated cycle runs when it exists.
What this is not
It is not an approval and not a rejection. It never edits the specification: an accuser who
rewrites the charge is not one. And it is not a second opinion — a finding either carries a
command somebody else can run, or it does not exist.
- "Every sentence checked out, so the spec is true" — a challenge reports what it could not test too; only what it managed to check reads as a clean bill of health.
Done when
Every checkable sentence has a verdict beside the command that produced it, the untested ones
are named, and the file is committed in the same branch as the specification it attacks.
