The Agent Skills Directory

[SAFE]: The skill defines a structured, logical workflow for resuming development work and contains no indicators of malicious intent or behavior.
[NO_CODE]: This skill consists entirely of markdown-based instructions and does not include any scripts, binaries, or automated package installations.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it instructs the agent to read potentially untrusted content from the codebase (e.g., TODOs, comments). \n
Ingestion points: Codebase, documentation, task lists, and recent edits (SKILL.md). \n
Boundary markers: Absent. \n
Capability inventory: File system writes and execution of verification tools. \n
Sanitization: Not specified in the instructions. \n
Note: This represents a standard risk for coding assistants; the skill's own logic is safe.

next-step