The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the user to provide an authenticated curl command containing sensitive credentials, including an apikey and a Bearer token. These secrets are processed by the local tool to manage session authentication.
[COMMAND_EXECUTION]: The instructions require the user to execute a local Python script (app.py) to start the ProcessFlow Review Tool at http://localhost:8765, which then manages local file state and network requests.
[INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data in the form of workshop transcripts and process definitions, creating a potential attack surface for indirect injection.
Ingestion points: Transcripts are pasted directly into the agent context; process data is fetched from the ProcessFlow API.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the transcripts are provided.
Capability inventory: The agent has the ability to execute shell commands, interact with a local web application, and perform write operations on production data through the Chrome MCP extension.
Sanitization: The skill does not define any sanitization or validation logic for the ingested transcript text.
[EXTERNAL_DOWNLOADS]: The underlying tool uses the system curl utility to fetch process information, scores, and reports from external API endpoints during the ingestion flow.

processflow-app