The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to offer and potentially execute remote shell and PowerShell scripts from an unverified third-party GitHub repository (colbymchenry/codegraph). This utilizes high-risk execution patterns that can lead to arbitrary code running on the system.
Evidence: curl -fsSL https://raw.githubusercontent.com/colbymchenry/codegraph/main/install.sh | sh in SKILL.md
Evidence: irm https://raw.githubusercontent.com/colbymchenry/codegraph/main/install.ps1 | iex in SKILL.md
[EXTERNAL_DOWNLOADS]: The skill recommends installing a global NPM package from an untrusted individual author.
Evidence: npm install -g @colbymchenry/codegraph in SKILL.md
[INDIRECT_PROMPT_INJECTION]: The skill analyzes local repository content and git diffs without providing boundary markers or sanitization, creating a vulnerability surface where malicious code comments could influence agent actions.
Ingestion points: codegraph index [path], git diff --name-only | codegraph affected --stdin in SKILL.md.
Boundary markers: Absent. No delimiters are used to wrap external code content.
Capability inventory: Shell command execution via codegraph CLI.
Sanitization: Absent.

codegraph-cli