Auto Repair Agent

AdvisoryAudited by Static analysis on May 10, 2026.

Overview

Detected: suspicious.dangerous_exec

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.

What this means

If the user or agent runs an unsafe command through watch mode, it could modify files, run programs, or perform other local actions.

Why it was flagged

The skill can execute a shell command supplied to its watch flow. This is documented and purpose-aligned, but it gives the invoked command the user's local permissions.

Skill content
const out = execSync(command, { encoding: 'utf8', timeout: 60000 });
Recommendation

Only use watch mode with commands you explicitly trust and intend to run; do not pass untrusted text directly into the command argument.

What this means

Bad learned patterns could lead to poor future repair advice, and sensitive error text may be stored locally if users teach it as a pattern.

Why it was flagged

Learned repair patterns are persisted locally and loaded in future runs, so incorrect or sensitive learned entries could influence later diagnoses.

Skill content
const PATTERNS_FILE = '.self-heal-patterns.json'; ... fs.writeFileSync(PATTERNS_FILE, JSON.stringify(custom, null, 2));
Recommendation

Review or delete .self-heal-patterns.json periodically, and avoid learning patterns that contain secrets, tokens, or private error logs.

What this means

Users have less provenance information and may not get an install-time warning that Node.js is needed.

Why it was flagged

The registry metadata gives limited provenance and does not declare the Node.js runtime that the SKILL.md examples require.

Skill content
Source: unknown; Homepage: none; Required binaries (all must exist): none; No install spec — this is an instruction-only skill.
Recommendation

Verify the publisher and review the included JavaScript before use, especially because the skill can run local commands.

Findings (1)

critical

suspicious.dangerous_exec

Location
self_healing_engine.js:292
Finding
Shell command execution detected (child_process).