The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external, potentially untrusted data from user arguments and git history to generate documentation updates in CLAUDE.md and docs/CODING_RULES.md.
Ingestion points: Data enters the agent context through the $ARGUMENTS parameter and the output of git diff and git log commands defined in SKILL.md.
Boundary markers: The instructions do not specify any delimiters or boundary markers to isolate external content from the system prompt.
Capability inventory: The skill has the capability to write to the local filesystem (CLAUDE.md, docs/CODING_RULES.md) and perform repository operations (git add, git commit).
Sanitization: There is no evidence of sanitization, filtering, or validation of the data retrieved from external sources before it is incorporated into the documentation updates.
[COMMAND_EXECUTION]: The skill uses Bash commands to perform environment checks, file system operations, and git repository interactions.
Evidence: Execution of test -f to check configuration files, wc -l for line counting, and git commands (diff, log, add, commit) for repository state management.

tuzi-update-claude-md