Triage
v1.0.0Auto-learns to prioritize tasks by urgency, impact, and user patterns. Grows smarter with each decision.
⭐ 2· 628·0 current·0 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
Name/description match the content: the SKILL.md plus patterns.md and signals.md define prioritization rules and learning behavior. There are no unrelated required binaries, env vars, or install steps.
Instruction Scope
Instructions direct the agent to observe incoming tasks, classify them, record overrides, generate pattern entries, and produce queue history/weekly summaries. This is coherent with triage, but it means the skill will collect message content and user corrections (potentially sensitive). The SKILL.md does not reference external endpoints or system paths, nor does it explicitly say how persistence is performed.
Install Mechanism
Instruction-only skill with no install spec or external downloads — low installation risk and nothing written to disk by an installer. All behavior comes from the runtime instructions.
Credentials
No environment variables, credentials, or config paths are required. The skill does not request unrelated secrets or system access.
Persistence & Privilege
Skill claims to 'auto-learn' and track overrides/patterns/queue history which implies persistent storage of user data. always is false (normal), but the SKILL.md does not specify where patterns/history are saved, who can read them, or retention policies — a privacy/visibility concern rather than an outright functional mismatch.
Assessment
This skill appears to do what it says: watch tasks, assign priorities, and learn from your corrections. Before installing, ask the developer or platform: (1) Where are learned patterns, overrides, and queue history stored (local skill memory, agent storage, or external service)? (2) Who can read or export that stored data? (3) How can you review, edit, or delete learned rules and history? (4) Can you restrict the skill to an observe-only or ask-before-applying mode while testing? If you handle sensitive messages, consider testing with non-sensitive items first and require explicit confirmation before any automatic pattern 'confirmation' is applied.Like a lobster shell, security has layers — review code before you run it.
latestvk97an4gxc0b2m75fkjm067p46s812v1y
License
MIT-0
Free to use, modify, and redistribute. No attribution required.