Clawdoc

Security checks across malware telemetry and agentic risk

Overview

Clawdoc appears to be a coherent local diagnostic tool, but it analyzes agent session history and can optionally save findings for future use.

Clawdoc looks appropriate for diagnosing OpenClaw sessions. Before installing or invoking it, remember that session logs may include private prompts, tool outputs, paths, and cost data; run it on the narrowest useful session path and review any output or .learnings content before sharing.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

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.

#
ASI02: Tool Misuse and Exploitation
Low
What this means

The skill may scan local OpenClaw session history and generate summaries about failures, token usage, costs, and tool behavior.

Why it was flagged

The skill tells the agent to run bundled local shell scripts over OpenClaw session logs. This is expected for the diagnostic purpose, but it is still local command execution over user data.

Skill content
Run: `bash {baseDir}/scripts/headline.sh ~/.openclaw/agents/main/sessions`
Recommendation

Run it only on session files or directories you intend to analyze, and prefer a specific session path when you want a narrow review.

#
ASI06: Memory and Context Poisoning
Medium
What this means

Diagnostic output may reveal private details from prior agent sessions if copied or shared.

Why it was flagged

Session JSONL files can contain prior user messages, assistant outputs, tool results, local paths, and cost metadata. The skill processes that retrieved context for diagnostics.

Skill content
Session JSONL files are the ground truth for all diagnostics
Recommendation

Review reports before sharing them, and avoid running cross-session analysis on logs that contain sensitive conversations unless needed.

#
ASI06: Memory and Context Poisoning
Low
What this means

If enabled, session-derived findings or potentially sensitive diagnostic context may persist and influence later agent behavior.

Why it was flagged

The skill documents an opt-in persistent write of diagnostic findings into a learnings file that future agents may reuse.

Skill content
To enable writing findings to `.learnings/LEARNINGS.md`, set `CLAWDOC_LEARNINGS=1` before running prescribe
Recommendation

Enable CLAWDOC_LEARNINGS only intentionally, and periodically review or clear .learnings/LEARNINGS.md.

#
ASI04: Agentic Supply Chain Vulnerabilities
Info
What this means

It may be harder to independently confirm that this package matches an upstream project or trusted release.

Why it was flagged

The registry metadata does not provide a source or homepage for provenance verification, even though the package contains executable scripts.

Skill content
Source: unknown
Homepage: none
Recommendation

Install from a trusted registry/source and compare with an upstream repository if provenance matters to your workflow.