The Agent Skills Directory

[COMMAND_EXECUTION]: The orchestration logic in main.js utilizes execSync to manage the workflow by executing shell scripts (collect.sh, publish.sh) and Python utilities located in the environment's script directories.
[EXTERNAL_DOWNLOADS]: The collect.sh script automates the retrieval of content from public platforms including Reddit, GitHub, and Hacker News using specialized Python scripts.
[DATA_EXFILTRATION]: The workflow performs automated data export by committing collected prompts and generated reports to a remote Git repository using git push operations.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting untrusted content from the web to generate new AI skills.
Ingestion points: Data is pulled from public forums and repositories via scripts/collect.sh.
Boundary markers: The orchestration layer does not implement explicit delimiters or safety instructions for the processed content.
Capability inventory: The skill has the ability to execute shell commands and publish data to the ClawdHub platform.
Sanitization: Content validation is performed during the conversion phase, including quality scoring, minimum length requirements, and verb presence checks as documented in the fix summary.

prompts-workflow