The Agent Skills Directory

[METADATA_POISONING]: The 'Skill Boundaries' section contains instructions and limitations that are entirely unrelated to the skill's stated purpose, specifically referencing 'audio production workflows' and 'audio engineering expertise' instead of content strategy. Additionally, there are duplicate section headers for 'Step 7,' indicating poor documentation quality or template reuse errors.
[INDIRECT_PROMPT_INJECTION]: The skill uses various prompt templates and markdown worksheets that interpolate user-provided text (such as audience descriptions, business topics, and content ideas) directly into the agent context without defensive measures. 1. Ingestion points: Prompt examples and 'Worksheet' markdown blocks in SKILL.md that take values like [audience], [topic], and [benefit]. 2. Boundary markers: Absent; there are no delimiters (e.g., XML tags or triple quotes) used to separate user data from instructions. 3. Capability inventory: None; the skill does not appear to use tools for file writing, network requests, or shell execution. 4. Sanitization: Absent; there is no validation or filtering of user input.
[EXTERNAL_DOWNLOADS]: The skill metadata references an external Node.js dependency via the '@clawfu/mcp-skills' package.

content-strategy