The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes various external data sources including design documents (e.g., plan.md, research.md), code diffs, container logs, and tool outputs. This creates a potential surface for indirect prompt injection where malicious instructions embedded in these files could attempt to influence the agent's behavior.
Ingestion points: Reads design files from 'docs/design//', research notes, and live container logs across all implementation phases.
Boundary markers: The skill does not define specific delimiters or "ignore embedded instructions" warnings when interpolating this external content into prompts for subagents.
Capability inventory: The skill utilizes powerful tools including 'Bash', 'Write', and 'Agent' to modify files, execute shell commands, and spawn autonomous agents.
Sanitization: There is no evidence of sanitization, filtering, or validation applied to external content before it is processed or used to guide subagent actions.
[COMMAND_EXECUTION]: The skill frequently uses the 'Bash' tool to execute development-related commands such as linting (ruff, pnpm), deployment management, and version control (gh, but). These operations are consistent with the skill's purpose of automating software engineering workflows and utilize standard development utilities.

implement-feature