The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes external data (Command Briefs and research files) to synthesize and generate new instruction sets (SKILL.md) for other AI agents. This creates a surface for indirect prompt injection where malicious instructions embedded in the input research could be propagated into the forged skill.
Ingestion points: Ingests data from local markdown files located in ai-tools/command-briefs/ and ai-tools/skills/*/research/.
Boundary markers: The instructions do not specify the use of delimiters or escaping when interpolating input data into generated templates.
Capability inventory: The skill is capable of scaffolding directory structures and writing markdown files; it does not execute the data it processes.
Sanitization: No automated sanitization is described, although the skill requires a manual 'Done Checklist' verification and grounding of all claims in research sources.
[EXTERNAL_DOWNLOADS]: The skill references and directs the agent to fetch documentation from https://cursor.com/docs/skills. These references target a well-known software development service and are used for legitimate instructional purposes.

weapon-forge