The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's methodology provides instructions for the agent to use curl for uploading visual evidence to Linear via signed URLs and utilizing git log to analyze project history. These are routine development tasks used here for legitimate project management purposes.
[DATA_EXFILTRATION]: The skill facilitates the transfer of project-related visual evidence to Linear. As Linear is a well-known project management service, this is considered functional behavior within the scope of the skill.
[PROMPT_INJECTION]: The skill directs the agent to install tracking protocols into persistent configuration files like CLAUDE.md, AGENTS.md, and .cursor/rules/. This is a standard mechanism for providing session context to agents and is used to maintain project tracking standards. The agent also processes repository content such as roadmaps and changelogs. Ingestion points: CLAUDE.md, AGENTS.md, roadmaps, and git history. Boundary markers: None explicitly defined. Capability inventory: Linear MCP tools, file system writes, and curl network access. Sanitization: The methodology includes a specific step (§4.8) to corroborate document-derived claims with the user before performing backfills.

linear-discipline