The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple bash commands including ls and git to collect repository state, commit history, and directory structures from ~/dev/claude and ~/dev/www-andrewriley-info.- [DATA_EXFILTRATION]: The skill performs git push operations to synchronize documentation updates to a remote repository. This behavior is intended for the skill's purpose but involves sending local project state to an external server.- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingest content from multiple files to perform its audit.
Ingestion points: Reads README.md, CLAUDE.md, PROFILE.md, and all SKILL.md files found in the skills/ directory.
Boundary markers: Absent. The skill does not use delimiters or explicit instructions to treat the ingested file content as untrusted data.
Capability inventory: The agent has the ability to execute shell commands, read files, edit files, and push changes to remote repositories.
Sanitization: None. The skill processes the raw text of the files directly to infer updates and communication patterns.

keep-current