The Agent Skills Directory

[SAFE]: Th e skill operates as a high-level instructional guide for structured reasoning and documentation workflows. It does not perform unauthori zed file access, network exfiltration, or command execution.- [PROMPT_INJECTIO N]: Th e skill is designed to ingest data from project-specific files such as docs/wiki and docs/backlog. While these represent an indirect prompt injection surface where content in those files could influence the agent's behavior, the risk is mitigated by the skill's advisory nature and the requirement for user confirmation before any final commands are executed.
Ingestion points: Reads files from docs/wiki/, docs/backlog/, and specific maintenance/ADR indexes.
Boundar y markers: The instructions do not define explicit del imiters for ingested file content.
Capability inventor y: Reads local project files; uses ask_user_question and web search tools.
Sanitizatio n: No explicit sanitization or filtering of ingested file content is mentioned.

discuss