Skip to content

ValentinNikolaev/agent-skills / agent-plugins/skills/wrap-agent-skill

wrap-agent-skill

Verified

Import, recover, normalize, or sync reusable skills into canonical sources and generated copies; not project memory.

FileKindSize
SKILL.mdskill md8.4 KB
agents/openai.yamlconfig242 B
references/import-review-rubric.mdmarkdown2.1 KB

references/import-review-rubric.md

View on GitHub
# Import review rubric

Score only from inspected source evidence. Cite the files behind material deductions.

## Hard blockers

Recommend `DO_NOT_ADD` regardless of score when any unresolved condition applies:

- redistribution rights are absent or incompatible;
- secrets, credentials, or private data cannot be removed safely;
- malicious or scope-expanding instructions remain;
- symlinks or paths escape the approved source root;
- destructive or production-side effects cannot be isolated;
- the source is not a reusable skill.

## Scored dimensions

| Dimension | Points | Full-credit anchor |
| --- | ---: | --- |
| Capability value | 0–20 | Solves a meaningful recurring task better than general instructions. |
| Reusability | 0–15 | Works across projects with explicit inputs and no hidden chat state. |
| Instruction clarity | 0–15 | Gives imperative workflow, boundaries, outputs, and validation. |
| Portability | 0–15 | Uses platform-neutral capabilities and parameterized paths. |
| Safety and trust | 0–15 | Has clear authorization, provenance, secret, side-effect, and failure boundaries. |
| Maintenance cost | 0–10 | Has a small essential surface and stable dependencies; award more for lower burden. |
| Validation readiness | 0–10 | Includes testable success criteria, safe script tests, or realistic evaluation inputs. |

Use integer scores. Explain any deduction of five or more points in a dimension.

## Recommendation thresholds

- `ADD`: 80–100, no hard blocker, and only minor normalization remains.
- `ADD_AFTER_IMPROVEMENTS`: 60–79, or 80+ with a material but fixable portability, safety, license, or validation gap.
- `DO_NOT_ADD`: 0–59 or any unresolved hard blocker.

## Decision report

Return:

1. total score and dimension breakdown;
2. strongest reasons to add;
3. risks, assumptions, and maintenance cost;
4. provenance and license result;
5. security and resource-inventory result;
6. recommendation enum;
7. exact normalization and validation plan;
8. any decision that requires user input.

Do not inflate the score because the source is popular, technically interesting, or already installed on one platform.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall