Back to skill

Security audit

商誉减值风险·逐条挂源

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious because scanners report network and credential-handling behavior that conflicts with its stated offline goodwill-analysis purpose; overall confidence is medium.

Review the network and credential behavior carefully before installing, especially the report that a key may be sent to a catalog-selected endpoint. The scan is incomplete: Skillspector marked analysis partial because some local references could not be resolved, even though all listed components were inspected; AIG completed with no findings. The suspicious verdict indicates real concerns for review, not a conclusion that the author acted maliciously.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (14)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · README.md (reported line 73)May include surrounding context.

bash
CATALOG="https://cuecue.cn/api/mcp-catalog"
DOMAIN_ID="$(curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
print(rows[0]["id"])')"
ROUTING_URL="$(curl -s "$CATALOG" | python3 -c 'import json,sys

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · README.md (reported line 76)May include surrounding context.

md
DOMAIN_ID="$(curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
print(rows[0]["id"])')"
ROUTING_URL="$(curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
print(rows[0]["routing"]["url"])')"
curl -s -X POST "$ROUTING_URL" \

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.en.md (reported line 156)May include surrounding context.

bash
CATALOG="https://cuecue.cn/api/mcp-catalog"
curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
[print(r.get("id"), r.get("external_status"), (r.get("routing") or {}).get("url", "-")) for r in rows]'
DOMAIN_ID="<replace with a domain id from the list above>"

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.en.md (reported line 160)May include surrounding context.

md
rows=json.load(sys.stdin)["data"]["domains"]
[print(r.get("id"), r.get("external_status"), (r.get("routing") or {}).get("url", "-")) for r in rows]'
DOMAIN_ID="<replace with a domain id from the list above>"
ROUTING_URL="$(curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
hit=[r["routing"]["url"] for r in rows if r.get("id")==sys.argv[1] and r.get("routing")]
print(*hit,sep="\n") if hit else sys.exit("not found: that id is not in the current catalog")' "$DOMAIN_ID")"

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 141)May include surrounding context.

bash
CATALOG="https://cuecue.cn/api/mcp-catalog"
curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
[print(r.get("id"), r.get("external_status"), (r.get("routing") or {}).get("url", "-")) for r in rows]'
DOMAIN_ID="<把这里换成上一行列出的那个域 id>"

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 145)May include surrounding context.

md
rows=json.load(sys.stdin)["data"]["domains"]
[print(r.get("id"), r.get("external_status"), (r.get("routing") or {}).get("url", "-")) for r in rows]'
DOMAIN_ID="<把这里换成上一行列出的那个域 id>"
ROUTING_URL="$(curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
hit=[r["routing"]["url"] for r in rows if r.get("id")==sys.argv[1] and r.get("routing")]
print(*hit,sep="\n") if hit else sys.exit("未取到:该 id 不在目录现值里")' "$DOMAIN_ID")"

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
100% confidence
Finding

The supplied code implements remote MCP service discovery and authenticated tool enumeration, not the declared goodwill-ledger workflow. Its curl requests directly contradict the zero-network claim. It contains no report extraction, source anchoring, goodwill accounting checks, or invocation of check_goodwill.py. Although the domain ID must be replaced for successful endpoint resolution, the catalog requests already demonstrate the mismatch.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
100% confidence
Finding

The supplied shell code performs remote service discovery and authenticated MCP tool listing, not document extraction, goodwill-ledger construction, or local validation. Its external HTTP requests directly contradict the declared offline boundary. The code also sends CUE_API_KEY to a remotely supplied routing URL, a credential-use behavior not disclosed in the declared purpose.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied shell code implements remote service discovery and authenticated MCP tool enumeration, not financial-document extraction or goodwill-ledger checks. It makes two requests to cuecue.cn and then posts to a routing URL returned by that catalog, attaching CUE_API_KEY as a bearer token. These operations materially differ from the stated offline purpose and are not supporting implementation details for the declared workflow. The domain ID remains a placeholder, so the final request requires a valid selection, but the catalog requests occur before that selection is validated.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
96% confidence
Finding

The example sends CUE_API_KEY to the first routing URL supplied by a remote catalog without validating its scheme or destination. A compromised or incorrectly configured catalog could direct the credential to an attacker-controlled endpoint or an unencrypted HTTP endpoint; the legitimate optional sourcing workflow does not eliminate this trust-boundary issue.

Content

Scanner excerpt · README.md (reported line 79)May include surrounding context.

md
ROUTING_URL="$(curl -s "$CATALOG" | python3 -c 'import json,sys
rows=json.load(sys.stdin)["data"]["domains"]
print(rows[0]["routing"]["url"])')"
curl -s -X POST "$ROUTING_URL" \
  -H "Authorization: Bearer $CUE_API_KEY" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' \

External Transmission

Medium
Category
Data Exfiltration
Confidence
96% confidence
Finding

The selected domain ID is checked against the catalog, but its routing URL is not checked against an independent trusted destination policy before receiving CUE_API_KEY. A compromised catalog could retain the expected ID while substituting an attacker-controlled or HTTP endpoint, exposing the credential despite the instructions not to print it.

Content

Scanner excerpt · SKILL.en.md (reported line 165)May include surrounding context.

md
hit=[r["routing"]["url"] for r in rows if r.get("id")==sys.argv[1] and r.get("routing")]
print(*hit,sep="\n") if hit else sys.exit("not found: that id is not in the current catalog")' "$DOMAIN_ID")"
[ -n "$ROUTING_URL" ] || { echo "lookup failed: ROUTING_URL is empty" >&2; exit 1; }
curl -s -X POST "$ROUTING_URL" \
  -H "Authorization: Bearer $CUE_API_KEY" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' \

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

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
96% confidence
Finding

CUE_API_KEY is attached to a URL taken directly from the remote catalog without independent HTTPS or origin validation. Matching the chosen domain ID and checking for a nonempty URL do not stop a compromised catalog from redirecting the credential to an unauthorized recipient.

Content

Scanner excerpt · SKILL.md (reported line 150)May include surrounding context.

md
hit=[r["routing"]["url"] for r in rows if r.get("id")==sys.argv[1] and r.get("routing")]
print(*hit,sep="\n") if hit else sys.exit("未取到:该 id 不在目录现值里")' "$DOMAIN_ID")"
[ -n "$ROUTING_URL" ] || { echo "取值失败:ROUTING_URL 为空" >&2; exit 1; }
curl -s -X POST "$ROUTING_URL" \
  -H "Authorization: Bearer $CUE_API_KEY" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' \

The missing-item reconciliation is documented as bidirectional but only checks from data rows to section six.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The checker only searches section six for text derived from missing data rows; it never verifies that every section-six entry corresponds to a missing row. Canonical missing-row placeholders also produce an empty tail and skip that search, allowing omissions or fabricated missing-item entries to evade the advertised reconciliation and undermine ledger integrity.

Content

No source excerpt is available for this finding.