The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several Git commands to manage the development workflow, including git status, git pull, git switch, git log, and git diff. Additionally, it includes steps to execute local linting, testing, and build processes as part of its verification phase.
[PROMPT_INJECTION]: The skill demonstrates an attack surface for indirect prompt injection through its reliance on external data from specification files to guide the agent's actions and determine which auxiliary skills are loaded.
Ingestion points: Local specification files located at context/specs/[unit-number]-[feature-name].md (Step 3).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between implementation details and potential instructions embedded within the spec files.
Capability inventory: The skill can execute Git commands, modify the file system (via the progress-tracker.md update and general implementation), and launch subagents with the task tool.
Sanitization: No validation or sanitization of the content within the specification files is performed before it is processed by the agent.

spec-impl