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
• The skill blindly installs a binary via curl/PowerShell from GitHub releases without checksum verification, which could be a supply-chain risk if the release is compromised. However, this follows common practice and is not a critical flaw.
• Indices de permissions à haut risque : exécution shell ou de commande
• Financial research output is not financial advice; require human review before any live investment decision
• The skill does not explicitly warn that CDP servers exposed on localhost can be accessed by other local processes, and the server itself may allow arbitrary browser automation from local users. This is typical for such tools and acceptable.