The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of skill names and descriptions provided by the user during the initialization phase.
Ingestion points: User-provided skill name and description used in SKILL.md generation and Slack messaging as described in SKILL.md.
Boundary markers: Not explicitly defined in the skill documentation to isolate user-provided content from agent instructions.
Capability inventory: Includes file system write access (creating directories, writing markdown files), file packaging (zip), and network access (Slack API via Rube integration) across the skill development pipeline.
Sanitization: No explicit sanitization or validation of user-provided content against instructional patterns is documented.
[COMMAND_EXECUTION]: The skill documentation indicates the use of Python 3.7+ scripts to automate the creation of skill structures and the validation of metadata.
[DATA_EXFILTRATION]: By design, the skill transmits information including skill metadata, descriptions, and direct file links to external Slack channels for team collaboration.

skill-share