Claude Skill

skill-prose-review

Use this skill when reviewing the contract completeness of Skills, Prompts, metadata, or QA documentation; triggers include Skill prose review, Prompt review, and contract audit.

LLM Mart · 0 points · 0 0 0

Virus-scanned Reviewed automatically before listing.

Full trust report

Download naodeng-awesome-qa-skills-skills_en_skill-engineering_skill-prose-review-c44b892.zip · 3 KB

naodeng/awesome-qa-skills 230 31 forks Updated 3d ago

Part of naodeng/awesome-qa-skills — 97 skills

Install

skills CLI npx skills add https://github.com/naodeng/awesome-qa-skills/tree/main/skills/en/skill-engineering/skill-prose-review

Claude Code

claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install naodeng-awesome-qa-skills@llmmart

Git git clone https://github.com/naodeng/awesome-qa-skills.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole naodeng/awesome-qa-skills collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Skill Prose Contract Review

When to use

Review a Skill, Prompt, metadata, example, or output format.
Check whether prose is executable, independently installable, and verifiable.

Workflow

Confirm scope and document role.
Trace trigger → input → rules → output → constraints → verification.
Report blockers, important suggestions, location, impact, and evidence.
List missing information and unverified claims.

Constraints

Do not invent behavior, tools, or facts.
Do not report static reading as runtime validation.
Do not rewrite files unless explicitly requested.

On-demand loading

Read prompts/skill-prose-review.md before producing the review.
Use evals/ for evaluation and report static and runtime evidence separately.

Final checklist

Scope and document role are explicit
Input, output, constraints, and verification are covered
Gaps and evidence levels are stated
No unauthorized file changes

Files (awesome-qa-skills)

agents

openai.yaml 378 B

version: 1
metadata:
  key: "skill-prose-review"
interface:
  display_name: "Skill Prose Contract Review"
  short_description: "Audit executable contracts, boundaries, and evidence requirements in Skills and Prompts."
  default_prompt: "Use the skill-prose-review skill to audit this Skill or Prompt contract."
policy:
  allow_implicit_invocation: true
schema_version: v1alpha1

evals

cases

basic-success.yaml 459 B

id: basic-success
title: Complete Skill contract review
description: Full review scope is provided.
input:
  prompt: Use skill-prose-review to review triggers, inputs, outputs, constraints, and verification. Report blockers, gaps, and evidence boundaries.
expect:
  must_contain: ["Blocking findings", "Information gaps", "Evidence boundary"]
judge:
  type: rule_based
  success:
    - output_contains:
        all: ["Blocking findings", "Evidence boundary"]

edge-incomplete-input.yaml 389 B

id: edge-incomplete-input
title: Missing review material
description: The user provides no target or runtime context.
input:
  prompt: Review whether this Skill passes, but no Skill content, scope, or runtime environment is provided.
expect:
  must_contain: ["Information gaps", "Not verified"]
judge:
  type: rule_based
  success:
    - output_contains:
        all: ["Information gaps"]

edge-scope-risk.yaml 348 B

id: edge-scope-risk
title: Unauthorized rewrite request
description: The user asks for an unrelated Skill to be rewritten.
input:
  prompt: Review Skill A, then rewrite all of Skill B without stating scope or evidence.
expect:
  must_contain: ["scope", "evidence"]
judge:
  type: rule_based
  success:
    - output_contains:
        all: ["scope"]

eval.yaml 709 B

version: 1
metadata:
  key: "skill-prose-review"
interface:
  display_name: "Skill Prose Contract Review"
  short_description: "Audit executable contracts, boundaries, and evidence requirements in Skills and Prompts."
  default_prompt: "Use the skill-prose-review skill to audit this Skill or Prompt contract."
policy:
  allow_implicit_invocation: true
schema_version: v1alpha1
environment:
  type: none
skills:
  - source: local_path
    path: .
engine:
  name: claude_code
cases:
  files: [evals/cases/basic-success.yaml, evals/cases/edge-incomplete-input.yaml, evals/cases/edge-scope-risk.yaml]
  defaults:
    timeout_seconds: 180
    max_turns: 8
    expect:
      exit_code: 0
report:
  formats: [json]

prompts

skill-prose-review.md 771 B

# Skill Prose Contract Review Prompt

You are a Skill engineering reviewer. Review only the requested scope; do not execute the business task described by the Skill or edit files.

Check triggers, inputs, assumptions, missing information, executable steps, observable outputs, constraints, refusal boundaries, evidence, bilingual consistency, and process-transcript residue.

Return:

```markdown
# Review conclusion
Conclusion: pass / conditional pass / fail

## Blocking findings
| Location | Finding | Impact | Evidence |

## Important suggestions
| Location | Suggestion | Reason |

## Information gaps
- …

## Evidence boundary
- Verified: …
- Not verified: …
```

Missing source files, runtime, or user goals are information gaps, not an invitation to guess.

SKILL.md 1.2 KB

---
name: skill-prose-review
description: Use this skill when reviewing the contract completeness of Skills, Prompts, metadata, or QA documentation; triggers include Skill prose review, Prompt review, and contract audit.
---

# Skill Prose Contract Review

## When to use

- Review a Skill, Prompt, metadata, example, or output format.
- Check whether prose is executable, independently installable, and verifiable.

## Workflow

1. Confirm scope and document role.
2. Trace trigger → input → rules → output → constraints → verification.
3. Report blockers, important suggestions, location, impact, and evidence.
4. List missing information and unverified claims.

## Constraints

- Do not invent behavior, tools, or facts.
- Do not report static reading as runtime validation.
- Do not rewrite files unless explicitly requested.

## On-demand loading

- Read `prompts/skill-prose-review.md` before producing the review.
- Use `evals/` for evaluation and report static and runtime evidence separately.

## Final checklist

- [ ] Scope and document role are explicit
- [ ] Input, output, constraints, and verification are covered
- [ ] Gaps and evidence levels are stated
- [ ] No unauthorized file changes