The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The LOG.md file (item #145) references the domain interface-design.dev as a visual design system model. This domain has been flagged as a phishing site by automated security scans. Since the skill's instructions for Flow 2 direct the agent to 'Study it at the source,' there is a risk that the agent or user may be directed to this malicious domain, potentially leading to credential theft or malware exposure.
[COMMAND_EXECUTION]: The skill uses the GitHub CLI (gh) and shell commands to fetch data from external repositories (gh issue list, gh api graphql, gh api repos/commits). While these are functional components of the triage tool, they involve the execution of commands that process data from arbitrary external sources.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its ingestion of untrusted external data.
Ingestion points: Content from GitHub issues, discussions, and commit messages from external repositories is read into the agent context in Flow 1 and Flow 2.
Boundary markers: The instructions do not define boundary markers or 'ignore' instructions for the external content.
Capability inventory: The skill uses the gh CLI and has the ability to read and write local repository files (e.g., LOG.md, CLAUDE.md).
Sanitization: No sanitization or validation of the fetched external content is described in the workflow, allowing malicious instructions embedded in issues or commits to potentially influence the agent.

source-triage