Installs into .claude/skills of the current project.
Are you the author of Evidence Gate?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/kontourai-evidence-gate)
---
name: "evidence-gate"
description: "Assess whether verification evidence and scope integrity support merge review. Produces a confidence report and Builder merge-readiness evidence."
---
# Evidence Gate
## Role
This is a Builder step skill.
Evidence Gate assesses whether the completed work has trustworthy, scoped
evidence for merge review. It does not fix code, publish a change, approve a
release, or deploy.
## Binding
| Context | Binding | Flow expectation |
| --- | --- | --- |
| Active Builder run | `builder.build` at `merge-ready` | `merge-readiness` |
| Standalone invocation | No Flow binding | No workflow mutation. |
For an active run, confirm the binding before evaluation:
The steering hook injects the run state at session start and re-injects it whenever it changes; if a steering state block appears anywhere in your context and you have not recorded evidence or a critique since it appeared, it is current. Run `flow-agents workflow status --session-dir <session-dir> --json` only after such a write or when no steering state block appears in your context at all; the write itself refuses an unbound, stale, or wrong-step binding either way.
Only a matching run may publish `merge-readiness`. A standalone confidence
report is local and must not call `workflow evidence`.
## Inputs
- The relevant verification and critique slices in `trust.bundle`.
- Changed-file scope from `RepositoryAdapter` or the execution record.
- Change and check records from `ChangeProvider` and `CheckProvider` when available.
- Known risks, accepted gaps, and repository policy evidence.
## Evaluation
Aggregate the existing canonical implementation scope, accepted criteria,
critique, observed commands, and available provider records. Reuse their exact
references and revision bindings. This step adds a confidence decision; it does
not require a second criterion ledger or another run of current checks.
1. Operate report-only. Do not patch source, alter checks, publish a change, or
rerun with weaker options to manufacture a passing decision.
2. Confirm acceptance evidence maps to the intended changed scope. Compare the
planned scope, actual diff, generated files, and provider-reported change.
Unexplained drift, weakened checks, or unrelated edits reduce confidence and
must be named.
3. Classify evidence by strength: direct runtime or test observation; provider
check/review record; reproducible command result; source inspection; or
unsupported assertion. Preserve command, revision, timestamp, scope, and
artifact provenance where available. Prose-only completion claims are not
acceptance proof.
4. Re-run only to resolve an actual stale result, changed scope, contradiction,
or missing required observation, and only when the command is safe and
reproducible. A stale,
unavailable, changed, or non-reproducible check is degraded to
`NOT_VERIFIED`; a previous pass must not remain current by assertion.
5. Produce `PASS`, `FAIL`, or `NOT_VERIFIED` for the confidence decision.
`NOT_VERIFIED` is the result for missing required proof, unavailable provider
information, or unresolved integrity questions.
6. On a matching active run, publish the decision through the public CLI:
```bash
flow-agents workflow evidence --session-dir <session-dir> \
--expectation merge-readiness \
--status <pass|fail|not_verified> \
--summary "Evidence confidence, scope integrity, and remaining risks are recorded." \
--evidence-ref-json '{"kind":"artifact","file":"<session-dir>/<slug>--evidence-gate.md","summary":"Evidence confidence, criterion coverage, and scope-integrity decision."}'
```
## Output
Record the confidence report as the `merge-readiness` slice in `trust.bundle`.
It contains the scope assessed, acceptance evidence, integrity findings,
evidence provenance, residual risks, `PASS`/`FAIL`/`NOT_VERIFIED` decision, and
recommended route.
Reuse the existing readable `Acceptance Evidence` table or canonical criterion
references when behavior is claimed. Add a table only when it resolves a scope
or evidence gap; do not copy the ledger merely to satisfy this step:
| AC id | Status | Command/Test Evidence | Source Evidence | Gaps |
| --- | --- | --- | --- | --- |
For a matching active run, publish that report as `merge-readiness`. This is a
merge-review confidence decision, not release or deployment authorization.