The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes git tools including git status, git add, and git commit to maintain development discipline and ensure atomic commits for each completed cycle. It also references npx for managing skill dependencies.
[EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install related components using npx skills add jwilger/agent-skills. This is a vendor-owned resource hosted under the author's namespace and is documented neutrally.
[INDIRECT_PROMPT_INJECTION]: The skill ingests project-specific documentation to provide context to subagents, creating a surface for potential indirect prompt injection.
Ingestion points: Reads and extracts content from docs/ARCHITECTURE.md, docs/glossary.md, and .factory/config.yaml.
Boundary markers: The orchestrator does not use explicit delimiters or warnings to isolate ingested project documentation from agent instructions.
Capability inventory: The skill possesses capabilities for file system modification via Write and Edit tools, as well as subagent delegation via the Agent tool.
Sanitization: External documentation content is interpolated directly into prompts without explicit sanitization or validation protocols.
[DYNAMIC_EXECUTION]: The orchestrator dynamically constructs and executes prompts for specialized subagents (RED, GREEN, DOMAIN, COMMIT) at runtime, incorporating gathered evidence and project-specific configuration.

tdd