The Agent Skills Directory

[SAFE]: No evidence of malicious instructions, prompt injections, or obfuscated content was found across the provided files. The skill focuses entirely on revenue operations and pipeline health diagnostics.
[DATA_EXPOSURE]: The skill is explicitly designed to handle sensitive CRM data, including deal amounts, contact details, and revenue forecasts. While there is no logic for unauthorized data exfiltration, the agent is inherently positioned to access high-value business information as part of its primary function.
[PROMPT_INJECTION]: Analysis of the instructions revealed no attempts to bypass safety filters, override system prompts, or extract internal configuration data.
[REMOTE_CODE_EXECUTION]: No external scripts, remote downloads, or package installation commands (such as curl, pip, or npm) were detected.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection because it is designed to ingest and process untrusted external data from CRM systems (e.g., deal names, contact engagement logs, and notes).
Ingestion points: The workflow involves pulling deal-level snapshots, contact activity, and MEDDPICC field data (SKILL.md).
Boundary markers: There are no explicit instructions for the agent to treat data as untrusted or to use delimiters to separate instructions from data.
Capability inventory: The skill is configured for analytical reporting and dashboard generation; no dangerous shell or code execution tools are requested in the configuration.
Sanitization: No explicit sanitization or validation of the input data is mentioned in the instructions.

agency-pipeline-analyst