The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon instructions contained in external specification files (User Stories, Technical Tasks, Work Items, and Test Cases). While this introduces a surface for indirect prompt injection, it is fundamental to the skill's purpose as an implementation agent.
Ingestion points: Specification files located in docs/specs/user-stories/, docs/specs/work-items/, and docs/specs/features/.
Boundary markers: The skill does not currently specify the use of delimiters or 'ignore' instructions when parsing these files.
Capability inventory: The agent has the ability to execute shell commands for testing, linting, and building, as well as performing Git operations and modifying local source files.
Sanitization: Instructions from specification files are interpreted directly by the model to guide code generation without explicit structural sanitization.
[COMMAND_EXECUTION]: The skill orchestrates typical developer tasks using shell commands. These include Git lifecycle management (branching, worktrees, commits) and project-specific verification tools (lint, build, unit/integration/E2E tests). All identified commands are consistent with the skill's primary implementation role.

work-implement