The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of official AI CLI tools from trusted organizations.
Evidence: npm install -g @anthropic-ai/claude-cli and pip install google-generativeai in SKILL.md.
[COMMAND_EXECUTION]: The skill uses local shell commands to execute AI queries and read files.
Evidence: Execution of claude, gemini, and cat commands throughout SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The content enhancement features read data from draft.md and interpolate it directly into AI prompts, which could allow malicious instructions within those files to manipulate agent behavior.
Ingestion points: draft.md (via shell variable ARTICLE) in SKILL.md.
Boundary markers: Absent; data is placed directly into the prompt string without delimiters.
Capability inventory: Includes network-capable CLI model invocation (gemini).
Sanitization: Absent; no escaping or filtering is applied to the file content before interpolation.

article