The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to execute a local JavaScript file and a repository analysis tool via the command line. • Evidence: In SKILL.md and references/plan-organization.md, the agent is instructed to run node .claude/scripts/set-active-plan.cjs {plan-dir} to track plan state. • Evidence: In references/research-phase.md, the agent is instructed to use repomix --remote <github-repo-url> for remote repository analysis.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection by ingesting and acting upon data from external or generated sources. • Ingestion points: The skill processes reports from researcher, scout, and debugger agents, and reads project-specific documentation such as ./docs/development-rules.md (documented in references/codebase-understanding.md and references/research-phase.md). • Boundary markers: No delimiters or specific instructions to ignore instructions within the ingested data are present. • Capability inventory: The agent can execute shell commands (node, repomix, gh), perform file system operations (creating directory structures and markdown files in the plans/ folder), and spawn additional agents. • Sanitization: The instructions do not specify any validation or sanitization for the content provided in reports or codebase files.

planning