The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from diverse external and internal sources, including market intelligence reports (LMTY), CRM opportunity data, call transcripts, and win/loss research as described in references/CONTEXT.md and references/INTERNAL.md.
Ingestion points: Data is retrieved from market intelligence tools and internal databases containing customer and sales team input (SKILL.md Context Acquisition section).
Boundary markers: The instructions in references/TRUST.md and references/INTERNAL.md require the agent to explicitly label sources, distinguish between facts and interpretations, and use specific templates to maintain clear data boundaries.
Capability inventory: The skill utilizes read-only tools to retrieve data from connected services and generates markdown-based battlecard artifacts. It does not possess network-write or system-modification capabilities.
Sanitization: The methodology requires the agent to quote verbatim from sources and label evidence by its representativeness (e.g., 'single account signal' vs 'repeated pattern'), which acts as a logical control against the propagation of malicious instructions embedded in the data.

competitive-battlecard