The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data from ClickUp API responses, which creates a potential surface for indirect prompt injection.
Ingestion points: ClickUp API response bodies are treated as sensitive content for evidence capture (references/official-docs.md).
Boundary markers: The instructions include explicit directives to "Redact or quarantine before sharing" evidence (SKILL.md).
Capability inventory: The skill utilizes Read, Glob, Grep, Write, and Edit tools to analyze repository configurations and perform containment or recovery actions (SKILL.md).
Sanitization: Robust sanitization instructions are provided, requiring the redaction of task text, comments, attachments, tokens, and webhook secrets from any output artifacts (references/official-docs.md, SKILL.md).
[DATA_EXFILTRATION]: The skill documentation includes strong warnings against the exposure of sensitive data.
The instructions explicitly command the agent to "never print the token, OAuth client secret, or webhook secret" (SKILL.md).
Operational guidance suggests using governed secret references rather than hardcoding credentials.
[EXTERNAL_DOWNLOADS]: The skill references official ClickUp documentation and sample code.
Links target official domains including developer.clickup.com and status.clickup.com, as well as the official clickup-APIv2-demo repository on GitHub (references/official-docs.md). These are well-known and legitimate resources for the service.

clickup-incident-runbook