The Agent Skills Directory

[DATA_EXFILTRATION]: The skill reads the repository's .git/config file to identify the remote origin URL. This is a standard operation required to configure GitHub or GitLab integrations. While .git/config can occasionally contain sensitive tokens, this access is directly associated with the skill's primary setup purpose.- [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection. Ingestion points: It reads repository files such as CLAUDE.md, AGENTS.md, and CONTEXT.md. Boundary markers: No explicit delimiters are defined for the ingested content. Capability inventory: The skill performs file writes to the repository and executes git commands. Sanitization: The skill mitigates risks by summarizing all findings and requiring explicit user confirmation before writing or updating any files.- [COMMAND_EXECUTION]: The skill executes the git CLI to inspect the repository state. It also generates templates that define standard workflows for gh and glab CLI tools, which is consistent with its intended functionality of establishing repository conventions.

setup-matt-pocock-skills