The Agent Skills Directory

[SAFE]: The skill is a collection of guidelines and principles for writing effective LLM prompts. It includes files on prompt mechanics, meta-prompting, and system prompt structure.
[NO_CODE]: No executable files, scripts, or command-line injections were found. The skill operates purely through natural language instructions provided to the agent.
[INDIRECT_PROMPT_INJECTION]: Evaluation of ingestion surface:
Ingestion points: The skill ingests untrusted user input through the <request> argument in SKILL.md.
Boundary markers: No explicit delimiters or boundary markers are defined for the input.
Capability inventory: No tools, file system operations, or network requests are defined in any file (SKILL.md, references/mechanics.md, references/metaprompting.md, or references/system-prompts.md).
Sanitization: No sanitization or escaping of the user input is performed.
Conclusion: Despite the ingestion surface, the lack of any functional capabilities prevents the possibility of malicious exploitation.
[CREDENTIALS_UNSAFE]: No API keys, tokens, or sensitive file paths (e.g., .env, .ssh) were found in the skill content.

prompt-engineering