Residue Classifier

AdvisoryAudited by Static analysis on May 6, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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.

NoteHigh Confidence

ASI06: Memory and Context Poisoning

What this means

The report may reveal private project structure, file names, and the absolute workspace path if the report is later shared or committed.

Why it was flagged

The script inventories all files under the current workspace and persists their relative paths in a local JSON report.

Skill content

for path in sorted(WORKSPACE.rglob('*')): ... rows.append({'path': rel, ...}) ... REPORT.write_text(json.dumps(out, indent=2) + '\n')

Recommendation

Run it only from the intended workspace, review `residue-classifier-report.json` before sharing, and delete or ignore the report if it contains sensitive file names.

NoteHigh Confidence

ASI09: Human-Agent Trust Exploitation

What this means

Users may interpret the 'critical' wording as stronger security or governance evidence than the simple filename-based classifier can prove.

Why it was flagged

The tool can label a workspace state as critical based on unmatched-file counts and then show a paid Stripe upgrade link.

Skill content

'critical residue anomaly detected; free classifier cannot safely resolve it' ... f'Residue Classifier Pro: {PRO_LINK}'

Recommendation

Treat the paid upgrade prompt as marketing, review the generated reasons yourself, and do not purchase or act solely because the report says 'critical.'