Back to skill

Security audit

小红书 / 抖音赛道日报 + 每日稿 (LinkDigest)

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it found a credential exposure risk and unclear disclosure about data sent to LinkDigest; the scan is incomplete.

Consider the suspicious verdict before installing. The main concern is that a configurable API destination could receive your key and submitted data; the scan also reports unclear disclosure about sending writing samples and account-post URLs to LinkDigest. Skillspector inspected only half of the components, and some instruction checks were partial, so the available evidence is incomplete.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Tainted flow: 'req' from os.environ.get (line 33, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Sending the API key as a bearer token to the intended LinkDigest API is legitimate, not inherently credential exfiltration. However, LINKDIGEST_BASE_URL accepts an arbitrary destination or plaintext HTTP endpoint, and the code attaches the real key without validating that destination; an attacker able to influence this configuration could receive the credential and submitted data. Default urllib redirect handling also lacks an explicit cross-origin credential boundary here.

Content

Scanner excerpt · scripts/watch.py (reported line 37)May include surrounding context.

python
"Authorization": "Bearer " + key(), "Content-Type": "application/json", "User-Agent": UA,
    })
    try:
        with urllib.request.urlopen(req, timeout=timeout) as r:
            return r.status, json.loads(r.read() or b"{}")
    except urllib.error.HTTPError as e:
        try:

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · scripts/watch.py (reported line 4)May include surrounding context.

python
#!/usr/bin/env python3
"""LinkDigest daily brief (每日稿) from the command line. Standard library only.

add / list / run / pause / resume / delete manage watches (POST/GET/PATCH/DELETE
/api/v1/watch); draft searches a topic once and writes a post from it
(/api/v1/search with digest + draft); checkup reads your own notes into a page
with a draft of the next one (/api/v1/digest/batch with synthesis + draft).

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

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.

The skill requests users' own writing samples and account-post URLs without disclosing that these inputs are sent to LinkDigest.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The skill solicits personal writing samples and account-post URLs without explicitly explaining that these inputs leave the local agent for third-party processing. The LinkDigest branding and API-key setup make remote processing somewhat apparent, but do not clearly disclose the transfer of potentially private samples or token-bearing URLs.

Content

No source excerpt is available for this finding.