The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/release_notes.py uses subprocess.check_output to execute PowerShell on Windows systems to retrieve persistent environment variables (GITCODE_TOKEN). While this is used for legitimate configuration retrieval, it involves executing shell commands at runtime.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It ingests commit messages from external GitCode repositories via the scripts/release_notes.py script and instructs the agent to 'summarize, categorize, and polish' this data. If a commit message contains malicious instructions, the agent might follow them during the summarization phase.
Ingestion points: Commit messages are ingested through the JSON output of the release_notes.py script (message, first_line, and body fields).
Boundary markers: Absent. There are no explicit delimiters or instructions telling the agent to treat the commit content as untrusted data.
Capability inventory: The agent has shell access to execute the Python extraction script.
Sanitization: The script performs basic filtering of 'noise' words (e.g., 'wip', 'log') but does not sanitize content against prompt injection attacks.

gitcode-release-notes