The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill establishes a vulnerability surface for indirect prompt injection by processing external data sources. 1. Ingestion points: SKILL.md identifies scraped web text, reviews, comments, emails, and third-party exports as untrusted inputs. 2. Boundary markers: SKILL.md includes explicit instructions to treat evidence as untrusted and to ignore any embedded instructions. 3. Capability inventory: references/connectors.md describes potential interaction with RevOps and marketing platforms such as HubSpot, Salesforce, and Google Ads. 4. Sanitization: The skill relies on instructional constraints rather than programmatic sanitization or filtering.
[NO_CODE]: The skill consists entirely of markdown instructions and reference materials. No executable scripts, binaries, or configuration files containing code were found in the provided files.

prospecting