The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) due to the ingestion of untrusted data from external platforms.
Ingestion points: The skill retrieves data from Supabase including {comment_snippet}, {job_posting_detail}, and {post_topic} which originate from third-party LinkedIn profiles and job listings.
Boundary markers: No boundary markers, delimiters (such as XML tags or triple backticks), or specific instructions are provided to help the agent distinguish between user instructions and untrusted data in templates like templates/sequence-templates/pain-language.md.
Capability inventory: The skill possesses the capability to write outreach logs to a database via client.log_outreach and generate local CSV files for campaign export.
Sanitization: There is no evidence of sanitization, escaping, or validation of the external lead data before it is interpolated into message prompts.
[COMMAND_EXECUTION]: The skill instructions specify the execution of a setup script (python3 tools/supabase/setup_database.py). The content of this script is not provided for analysis, preventing verification of its security behavior.

linkedin-outreach