The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and classify untrusted external inputs in the form of prospect replies (email and LinkedIn) as described in REPLIES.md.
Ingestion points: Incoming messages are analyzed by the agent to determine the appropriate response category (e.g., referral, objection, or information request).
Boundary markers: The instructions do not specify the use of prompt boundary markers or explicit directives to ignore instructions that might be embedded in the reply text.
Capability inventory: The agent has the capability to draft and, in specific cases (out-of-office or unsubscribe), autonomously send replies, which could be exploited if an incoming message contains malicious instructions.
Sanitization: No sanitization or filtering of incoming message content is mentioned before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill references several official documentation sources and public repositories for guidance.
Fetches compliance and deliverability rules from official sources including Google Support, the Federal Trade Commission (FTC), the Federal Communications Commission (FCC), and LinkedIn's User Agreement.
References marketing frameworks and benchmarks from the public coreyhaines31/marketingskills repository.

cold-outreach