The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data from social media accounts to extract structural patterns for post drafting, creating a vector for indirect prompt injection.
Ingestion points: The skill retrieves recent posts from external 'inspiration accounts' (X and LinkedIn) using tools like social-fetch or agent-browser as described in SKILL.md (Step 4) and references/inspiration.md.
Boundary markers: The instructions lack explicit boundary markers or directives to the agent to ignore potentially malicious instructions embedded within the scraped social media content.
Capability inventory: The skill possesses significant capabilities, including the ability to create and syndicate social media drafts to the user's personal X and LinkedIn accounts via the mcp__typefully__typefully_create_draft tool.
Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed by the LLM to influence the generated drafts. However, the skill does include human-in-the-loop checkpoints, requiring user approval before drafts are created and scheduled.

jab-hook