The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads and processes the content of other skills (specifically their SKILL.md and test files) to automate compliance improvements. This creates a surface for indirect prompt injection where a target skill's content could attempt to subvert the improvement process.
Ingestion points: Reads SKILL.md, unit.test.ts, and triggers.test.ts from the plugin/skills/ directory (SKILL.md, LOOP.md).
Boundary markers: Absent. The instructions do not specify any delimiters or warnings to treat the loaded content as potentially untrusted data.
Capability inventory: The skill can write to the repository via 'git commit' and execute local project scripts via 'npm test' and 'npm run tokens'.
Sanitization: Absent. The skill does not perform validation or sanitization of the ingested content before using it to generate updates.
[COMMAND_EXECUTION]: The skill uses several standard CLI tools to perform its intended tasks, including 'git' for version control, 'npm' for executing tests and token analysis, and 'cp' for scaffolding test directories. These commands are constrained to the local environment and the project's directory structure.
[DYNAMIC_EXECUTION]: The skill creates new test files by copying existing templates and then executes these files using 'npm test'. This automated test generation is a primary feature of the skill and represents a low-risk use of dynamic execution for quality assurance purposes.

sensei