Back to skill

Security audit

概念雷达 Concept Radar

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it reports a possible exposure of a GitHub token through unvalidated URLs or redirects, alongside undeclared tool capabilities; the review is based on incomplete evidence.

Treat this as a review-needed skill, not a finding that its author acted maliciously. Before installing, consider whether you need it and whether you are comfortable with its reported access; avoid making a GitHub token available until the token-handling concern is resolved. The scan was partial because some referenced paths were missing, so its conclusions are based on incomplete evidence.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_read, file_write, network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/concept_radar.py (reported line 528)May include surrounding context.

python
since = parse_time(cursor) - dt.timedelta(seconds=OVERLAP_SECONDS)
        params["since"] = since.isoformat().replace("+00:00", "Z")
    url = (
        f"https://api.github.com/users/{urllib.parse.quote(username)}/gists"
        + urllib.parse.urlencode(params)
    )
    request_key = canonicalize_url(url)

The GitHub scan silently reads and uses GITHUB_TOKEN when it is present.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

Using GITHUB_TOKEN for GitHub authentication is conventional, but the surrounding implementation does not constrain where the resulting Authorization header can travel. Authenticated headers are reused for response-supplied pagination and detail URLs, and the default urllib redirect handler can preserve Authorization across origins; an attacker-controlled destination reached through these mechanisms could receive the token, although control of such a response or redirect is a prerequisite.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · tests/test_concept_radar.py (reported line 131)May include surrounding context.

python
def opener(request, timeout=30):
            url = request.full_url
            calls.append(url)
            if url.startswith("https://api.github.com/"):
                raise urllib.error.HTTPError(
                    url, 403, "rate limited", {"x-ratelimit-remaining": "0"},
                    io.BytesIO(b'{"message":"API rate limit exceeded"}'),

Digest rendering fixes user-facing headings and labels to Chinese without a language choice.

Low
Category
Output Handling
Confidence
95% confidence
Finding

The generated digest unconditionally uses Chinese headings and labels, regardless of the user's language. No locale option, opt-in, or region-specific justification is present in the provided file.

Content

No source excerpt is available for this finding.