Plan de validation

Ce que l’Agent doit faire ensuite

1. 1Inspect repository, README/SKILL.md, license, and recent commits before production use.
2. 2Install in an isolated workspace or sandbox with no production secrets available.
3. 3Run the smallest representative task and record files touched, commands run, network access, and outputs.
4. 4Compare the selected skill against at least one alternative when the eval status is review or failed.
5. 5Promote only after the agent reports a successful verification result and unresolved warnings are accepted.

Ne pas utiliser quand

Conditions exigeant une autre skill

• Équipes qui nécessitent un SLA soutenu par le fournisseur
• production agents without a repository review
• SKILL.md only provides a high-level quick reference with rule names and categories, lacking the detailed explanations, code examples, and workflow steps needed for an agent to correctly apply the guidelines. It is effectively a summary rather than a self-contained skill.
• Indices de permissions à haut risque : Secrets or environment access
• Permission surface may require sandboxing
• Inconsistency in rule count: the description in SKILL.md states 'Contains 62 rules' while AGENTS.md and README mention '40+ rules'. This discrepancy undermines clarity.