The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it fetches and processes data from external websites and job postings to generate scores and sales talk tracks.
Ingestion points: Uses web_search and fetch_webpage to monitor LinkedIn profiles, job boards, and news sites for growth signals.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore commands that might be embedded in the retrieved web content.
Capability inventory: The agent possesses the capability to write reports to the local file system and generate high-stakes sales communication based on the input data.
Sanitization: No sanitization or filtering mechanisms are described for the external data before it is presented to the LLM for reasoning.
[DATA_EXFILTRATION]: The skill processes sensitive business and financial data, creating a risk surface for data exposure.
The skill is designed to ingest customer lists containing sensitive fields such as MRR/ARR, seat usage, and primary contact LinkedIn URLs.
While this data processing is central to the skill's primary function, the combination of sensitive data access and network-enabled tools (web_search) represents a potential exfiltration path if the agent is manipulated via prompt injection.

expansion-signal-spotter