The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands like echo, date, and git to create and update a local .implementation-log.md file. It also includes a shell script template for a git post-commit hook to automate the logging process.
[INDIRECT_PROMPT_INJECTION]: The skill provides a mechanism to ingest external data from git commit messages into log files, which may be subsequently read by the agent. This creates a potential vulnerability surface where malicious instructions could be embedded in commit history.
Ingestion points: Git commit messages accessed via git log -1 --pretty=%B in the post-commit hook template within SKILL.md.
Boundary markers: The log template does not use specific delimiters or warnings to separate untrusted commit content from the rest of the documentation.
Capability inventory: The skill performs local filesystem writes through shell redirection (>>) in SKILL.md.
Sanitization: The skill does not provide methods for sanitizing or escaping commit message content before it is stored in the log file.

implementation-logger