The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources which could be used to influence agent behavior.
Ingestion points: User task descriptions (passed directly to the --intent flag in SKILL.md), automated pipeline findings (the findings table in the gate output), and repository files such as .gitignore or configuration YAMLs.
Boundary markers: The instructions lack explicit delimiters or safety warnings to ignore instructions embedded in the user's intent or the tool's findings.
Capability inventory: The agent has the authority to modify files, execute Git operations (commit, branch, push), and trigger pull request creation through the no-mistakes CLI.
Sanitization: There are no instructions for the agent to sanitize or escape user-provided intent strings or findings descriptions before processing them or passing them to CLI commands.
[COMMAND_EXECUTION]: The skill relies on the execution of external CLI tools to perform its core functions.
Evidence: Instructs the agent to execute git (status, commit, branch) and the no-mistakes tool suite (subcommands: axi run, axi respond, axi status, doctor, init, sync).

no-mistakes