The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it retrieves and processes untrusted data from the web and social media. This data is then used to influence agent tasks like lead scoring and outreach generation.
Ingestion points: The enrichment-agent.md and signal-scorer.md agents retrieve external data from the X API (tweets, bios) and web search results (news, articles).
Boundary markers: The instructions do not define clear delimiters or "ignore" directives to separate retrieved data from the agent's core operational logic.
Capability inventory: The agents possess powerful tools, including Bash, WebFetch, and WebSearch, which represent a significant attack surface if an attacker embeds malicious instructions in public social profiles or posts.
Sanitization: There is no evidence of input validation or content filtering to prevent retrieved strings from potentially triggering unintended agent behavior.

lead-intelligence