The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The file diagnostics.md contains a hardcoded PostHog project API key used for capturing telemetry data.
[COMMAND_EXECUTION]: The skill uses various shell commands to perform repository operations (git), run build and test scripts (pnpm), and manage system services (launchctl).
[DATA_EXFILTRATION]: When opted-in by the user, the skill collects system metadata (version, OS, architecture) and transmits it to a third-party analytics service (PostHog).
[EXTERNAL_DOWNLOADS]: The skill fetches updates and configurations from the official project repository hosted on GitHub.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external content from the repository's CHANGELOG.md.
Ingestion points: The skill analyzes diffs of the CHANGELOG.md file after an update.
Boundary markers: No delimiters or explicit instructions are provided to the agent to ignore potentially malicious content within the changelog.
Capability inventory: The skill can invoke additional skills using the Skill tool based on the content of the parsed changelog entries.
Sanitization: There is no validation or sanitization performed on the skill names identified in the changelog before they are executed.

update-nanoclaw