The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Git commands via subprocess.run to retrieve repository metadata and commit details. Analysis of scripts/append_git_push.py confirms that commands are executed using argument lists with shell=False, which effectively mitigates shell injection risks.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided text for work descriptions. Ingestion occurs via the --content argument in scripts/append_entry.py. While explicit prompt boundary markers are absent, the skill provides a capability inventory limited to local file writes and Git metadata retrieval. Robust sanitization is implemented in scripts/worklog_core.py via the sanitize_field function, which prevents log injection by stripping newlines and field delimiters.
[SAFE]: The skill follows security best practices for local data management, specifically using os.open with mode 0o600 to ensure worklog files are only accessible by the owner. It also includes an explicit hostname allowlist to restrict Git recording to trusted domains.

weekly-worklog