The Agent Skills Directory

[DATA_EXFILTRATION]: The skill processes sensitive business and customer information, including ARR/MRR data, customer lists, and contact details. The combination of access to this internal data and the use of tools with network capabilities like web_search and fetch_webpage presents a risk of data exposure or exfiltration.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the public web to generate summaries and talk tracks.
Ingestion points: External company websites, career pages, and news sources fetched via web_search and fetch_webpage in SKILL.md.
Boundary markers: Absent; there are no instructions provided to the agent to treat fetched data as untrusted or to ignore embedded instructions.
Capability inventory: The skill has file-write permissions (saving reports to the local directory) and network access via tools.
Sanitization: Absent; the instructions do not include any steps for sanitizing or validating external content before it is processed by the LLM.
[COMMAND_EXECUTION]: The skill documentation suggests setting up a cron job that executes a local Python script (run_skill.py), which involves executing code on the host machine to automate the monitoring tasks.

expansion-signal-spotter