The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests untrusted data from the codebase (comments and file contents) and uses that data to generate or update a register file.
Ingestion points: Scans codebase files for TODO, FIXME, HACK, and @deprecated markers (Phase 2A) and reads the docs/tech-debt-register.md file (Phase 2C).
Boundary markers: The instructions do not specify the use of delimiters or clear separation between the scanned code content and the agent's operational instructions.
Capability inventory: The skill is permitted to use Read, Glob, Grep, and Write tools to access and modify the repository.
Sanitization: The instructions do not describe any sanitization, escaping, or filtering of the content extracted from codebase comments before processing or writing it to the technical debt register.

tech-debt