Agent-design-language repo-review-docs

Specialist docs reviewer for a multi-agent repository review. Use when a review packet needs a bounded documentation role focused on misleading docs, stale commands, onboarding gaps, release and demo truth drift, API/CLI contract drift, and whether documentation overclaims behavior without editing docs.

install

source · Clone the upstream repo

git clone https://github.com/danielbaustin/agent-design-language

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/danielbaustin/agent-design-language "$T" && mkdir -p ~/.claude/skills && cp -r "$T/adl/tools/skills/repo-review-docs" ~/.claude/skills/danielbaustin-agent-design-language-repo-review-docs && rm -rf "$T"

manifest: adl/tools/skills/repo-review-docs/SKILL.md

source content

Repo Review Docs

Review documentation truth and usability as one specialist in the multi-agent repo review suite.

This skill focuses on docs as operational surfaces. It is not a copyediting role unless copy clarity affects real user or reviewer behavior.

Quick Start

Confirm the repo, branch, path, diff, or review packet scope.
Identify docs that make operational claims: README, milestone docs, CLI commands, release notes, demos, schemas, and onboarding guides.
Compare docs against repo-visible commands, files, and behavior.
Emit findings first, with stale command or truth-drift evidence.
Hand the artifact to
```
repo-review-synthesis
```
for cross-role assembly.

Focus

Prioritize:

stale commands and broken paths
docs that claim behavior not present in code, tests, or artifacts
onboarding gaps that prevent a reviewer or operator from reproducing work
demo, release, milestone, and closeout truth drift
API, CLI, schema, or skill-contract documentation drift
ambiguity that can cause unsafe operation or wrong workflow execution

Defer primary ownership of these areas to other specialists:

executable defects:
```
repo-review-code
```
security exploitability:
```
repo-review-security
```
missing tests:
```
repo-review-tests
```
final dedupe and ordering:
```
repo-review-synthesis
```

Required Inputs

At minimum, gather:

```
repo_root
```

one concrete target:

```
target.target_path
```
```
target.branch
```
```
target.diff_base
```
```
target.review_packet_path
```

Useful additional inputs:

```
changed_paths
```
```
claimed_commands
```
```
demo_docs
```
```
release_docs
```
```
validation_mode
```

If there is no concrete repo or slice target, stop and report

blocked

Output Expectations

Default output should include:

findings first
reviewed docs surfaces
commands or claims checked
validation performed or not run
residual docs risk

Use the shared suite contract in

adl/tools/skills/docs/MULTI_AGENT_REPO_REVIEW_SKILL_SUITE.md

when ADL expects a structured artifact.

Stop Boundary

Stop after producing the docs-review artifact.

Do not:

rewrite documentation
spend the review on style nits when truth drift exists
claim demo/release readiness without proof
hide docs findings that affect operator safety or reviewer reproducibility