The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the ! syntax in the SKILL.md frontmatter to execute git diff --stat HEAD~20 at load time. It also instructs the agent to read the git log. These are standard, legitimate operations used to provide necessary context for the skill's scope analysis purpose.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection. 1. Ingestion points: The skill reads design documents in design/gdd/, sprint plans in production/sprints/, milestone definitions in production/milestones/, and the repository's git log. 2. Boundary markers: Absent. 3. Capability inventory: The skill utilizes the Read, Glob, and Grep tools as defined in the allowed-tools section of SKILL.md. 4. Sanitization: Absent. Malicious instructions embedded in the project files or commit messages could attempt to bias the risk assessment or influence the cut/defer recommendations.
[DATA_EXFILTRATION]: The skill accesses sensitive internal project documentation and the codebase. However, there are no network-capable tools or remote code execution patterns present, ensuring the data remains local to the execution environment.

scope-check