Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
updated to 1.5 version, 100% fit for clawhub skill format
What This Skill Does
Enforces evidence-first verification before answering claims about system state, files, configs, connectivity, and model selection. Each claim is classified as Verified, Inferred, or Unknown based on direct evidence from tools, files, or command outputs. Assertions without supporting evidence are never upgraded to confirmed status.
Prevents confident but wrong answers by requiring primary evidence such as files, logs, and command outputs rather than memory or plausible assumptions.
When to use it
- Confirming which AI model is actually configured before debugging unexpected behavior
- Verifying a service is running before chasing connection errors
- Checking whether a config change was saved and applied after editing a file
- Validating that a command ran and completed before reporting success
- Determining if a volume or directory is mounted before diagnosing missing files
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Is my OPENAI_API_KEY set in the environment?
- 1Extract the claim: OPENAI_API_KEY is present in the current environment
- 2Run an evidence command to inspect environment variables directly
- 3Classify the result as Verified if the key is found, Unknown if access is restricted or absent
- 4Cite the specific output line or file path that supports the classification
- 5Report the finding with next-step commands for any Unknown claims
Verified: OPENAI_API_KEY is set (found in .env, line 3), or Unknown: variable not present in current shell — run `printenv OPENAI_API_KEY` to confirm