The Agent Skills Directory

[SAFE]: The skill is entirely descriptive, providing a framework for task decomposition and decision management on an issue tracker. It does not include scripts, binaries, or complex configuration files.
[SAFE]: The skill uses disable-model-invocation: true in its frontmatter, preventing the model from autonomously executing commands without user intervention, which is a strong safety posture for a planning-centric skill.
[SAFE]: Interaction with the issue tracker is mediated through external sub-agents (e.g., /research, /prototype, /grilling), ensuring that the primary skill remains scoped to the planning logic rather than possessing direct, unconstrained access to external APIs.
[SAFE]: No obfuscation, hardcoded credentials, persistence mechanisms, or malicious prompt injection patterns were found in the markdown instructions or the policy configuration.

wayfinder