v1.0.2

Stress Toolkit

ReviewClawScan verdict for this skill. Analyzed May 1, 2026, 8:35 AM.

Analysis

This is mostly an offline relaxation guide, but it overstates its crisis-detection safety behavior, especially for the English and moderate-risk cases it advertises.

GuidanceUse this only as a simple offline set of relaxation exercises. Do not rely on its crisis detection; verify or improve keyword coverage, moderate-risk handling, and crisis referral wording before depending on it for safety-sensitive mental-health situations.

Findings (1)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Abnormal behavior control

Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.

Human-Agent Trust Exploitation

SeverityMediumConfidenceHighStatusConcern

SKILL.md

"High-Risk Detection ... Keywords: suicide, self-harm, don't want to live..." and "Moderate-Risk Detection ... Response: Express understanding and care"

The artifact advertises crisis-safety behavior that is not matched by the included implementation: scripts/crisis_detector.py lists Chinese-only risk terms, and scripts/handler.py only special-cases risk_level == "high" before falling through to normal techniques. This can make users or agents over-trust crisis handling.

User impactA user in serious distress, especially using English phrases or moderate-risk language, might not receive the promised referral or extra-care response and may be shown ordinary relaxation guidance instead.

RecommendationTreat the skill as a basic relaxation reference only. Update the detector and tests to match the documented languages and moderate-risk handling, and clearly state language, locale, and emergency-resource limits.