The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted user-provided data, including sales pipeline definitions, CRM stage history, and deal samples. This creates a theoretical surface for indirect prompt injection if malicious instructions were embedded in the pasted pipeline data.
Ingestion points: Data enters the agent's context through the 'Interview' workflow described in SKILL.md and the 'Artifact checklist' in references/audit-intake.md.
Boundary markers: The instructions do not explicitly define delimiters or 'ignore' warnings for the untrusted data, though the workflow mandates a structured analytical approach against established frameworks.
Capability inventory: The provided files contain no executable code, subprocess calls, file-system writes, or network operations.
Sanitization: The skill relies on the agent's internal reasoning and alignment with the provided frameworks (Gartner, GitLab) to process the data.
[NO_CODE]: The skill consists entirely of markdown instructions and documentation. No Python, Node.js, or shell scripts are included, which significantly limits the potential for technical exploitation.
[SAFE]: The skill uses established industry frameworks and does not exhibit any signs of obfuscation, exfiltration, or persistence. All external references are internal to the author's skill ecosystem (mbfinotti).

pipeline-stage-definition-audit