Skip to content
Agentmart
Submit a tool

evidence-gate

Gate completion claims by mapping every observable goal or acceptance criterion to typed command, artifact, Git, GitHub, or URL evidence.

Skill by cheshirecode

IndexedUnlicense
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/cheshirecode-evidence-gate

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Gate completion claims by mapping every observable goal or acceptance criterion to typed command, artifact, Git, GitHub, or URL evidence. Use before marking a multi-clause task, agent loop, deployment, PR, or verification workflow complete, especially when tests passing does not prove delivery, merge, or user-visible success.

Use cases: productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found
Runs as
Python package

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/evidence-gate

What's inside

Skill name
evidence-gate
Description
Gate completion claims by mapping every observable goal or acceptance criterion to typed command, artifact, Git, GitHub, or URL evidence. Use before marking a multi-clause task, agent loop, deployment, PR, or verification workflow complete, especially when tests passing does not prove delivery, merge, or user-visible success.
License
Unlicense
Instructions
47 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Files (7)

  • SKILL.md
  • openai.yaml · agents/
  • health-evidence.md · references/
  • recording.md · references/
  • evidence_gate.py · scripts/
  • test_cli_contract.py · tests/
  • test_evidence_gate.py · tests/

SKILL.md (first lines)

# evidence-gate

Require evidence coverage, not a persuasive completion summary. The script
checks that every declared criterion has at least one typed evidence record; the
agent remains responsible for verifying that each record is truthful and
relevant.

Skip when one observable outcome has one sufficient check.

## Resolve the skill directory

Resolve `<skill-dir>` to the directory containing this `SKILL.md`. In most
agent contexts, this is the path from which the skill was loaded. If uncertain,
search for `evidence_gate.py` under the skill root.
`loop-engineering/references/resolvers.md` owns this resolver pattern, including the per-host variants and the fixture that executes them; the line below is the same pattern with this skill's own sentinel file.

```bash
# Roots checked in order; empty when absent — never a bogus "./..":
SKILL_DIR="$(for r in ~/.claude/skills ~/.agents/skills ~/.cursor/skills ./skills; do f=$(find -L "$r" -name evidence_gate.py -print -quit 2>/dev/null); [ -n "$f" ] && { dirname "$(dirname "$f")"; break; }; done)"
```

All script invocations below use `python3 <skill-dir>/scripts/evidence_gate.py`.

## Record and check coverage

Read [references/recording.md](references/recording.md) when declaring criteria,
recording evidence, or checking completion; it owns the CLI examples and exit codes.
Start with `python3 <skill-dir>/scripts/evidence_gate.py init`.

**Evidence kinds:** `command`, `artifact`, `git`, `github`, `url`. Never
record model prose as evidence. Evidence of a change is not evidence of its
health — see [references/health-evidence.md](references/health-evidence.md).

## Inspect the gate

Run `python3 <skill-dir>/scripts/evidence_gate.py show --gate <gate-file>` to
print the full gate JSON (all criteria and their evidence records). Useful for
debugging or reviewing what has been recorded.

## Digest semantics

The SHA-256 digest proves which coverage artifact was checked, not that an
evidence source was interpreted correctly. Preserve the gate file with the task
evidence. If the gate file is modified after `check`, the digest no longer
matches; re-run `check` to regenerate it.

From cheshirecode/dotfiles, Unlicense.

More from cheshirecode/dotfiles

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 12, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Unlicense
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from cheshirecode

Cite this page

evidence-gate skill. Agent Mart. https://agentmart.dev/skills/cheshirecode-evidence-gate (last updated 2026-10-24)