Install
$ agentstack add skill-corrame-skill-refinement-mature-skill ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Mature Code Writer Skill
Use this for focused implementation work where the spec is bounded, ownership is clear, and the result can be verified with tests or equivalent checks.
Method
- Restate the spec, owned files, forbidden scope, and verification target.
- Read the smallest source and test slices that can prove the change.
- Inspect current signatures, fixtures, and runtime entry points before writing
calls or patches.
- Make the minimal change that satisfies the spec.
- Run the closest relevant check first.
- Attribute failures before changing more code.
- Escalate if the fix requires product, architecture, or ownership decisions.
- Report changed files, checks run, failures handled, residual risk, and one
lesson candidate if the task exposed a reusable pattern.
Avoid
- Expanding scope because adjacent code looks messy.
- Committing from a worker context.
- Treating static checks as a substitute for behavior tests when behavior tests
exist.
- Rewriting tests to match accidental behavior before confirming intended
behavior.
- Suppressing import, fixture, or collection failures without explaining the
underlying dependency or wiring issue.
Output
Spec:
Files changed:
Validation:
Failures handled:
Residual risk:
Lesson candidate:
Local Knowledge
This section contains rules promoted from repeated implementation tasks. They are examples of project-shaped knowledge that should not be guessed from a fresh prompt alone.
- Run local validation with
.venv/bin/pythonunless the project explicitly
documents another interpreter path. Shell python may point outside the project environment.
- Before adding or updating a call site, inspect the live source for the target
function's current signature and keyword-only arguments. Do not rely on memory or stale documentation.
- When changing runtime paths, entry points, dependency wiring, or environment
loading, run the full test suite after focused tests pass. These changes can break modules that focused tests do not import.
- If test collection fails because an optional dependency is missing, decide
whether the dependency belongs in dev extras or whether the import should be lazy. Do not hide the failure by skipping collection.
- When path behavior becomes profile-aware, tests that rely on hardcoded
repo-root tmp_path fixtures may silently test the wrong semantics. Pass an explicit root into affected fixtures.
- If a focused fix fails three times in the same area, stop widening the patch
and report the failure boundary. Repeated local edits can turn a small bug fix into an accidental refactor.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Corrame
- Source: Corrame/skill-refinement
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.