Back to skill

Security audit

企业公告债务画像

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed read-only public-notice lookup tool, but it also creates and persists guest access tokens through local shell scripts in a way users should review carefully.

Install only if you are comfortable with this skill running local bash/python helpers, contacting agent.debtop.com, and storing a short-lived guest refresh token plus device identifier in your user cache. Prefer a platform-managed MCP connector/OAuth path when available, and do not pass or allow an alternate BASE URL unless you control and trust that service.

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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (25)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

Including proof-of-work solving and SHA-256 nonce brute-force behavior is outside the expected scope of a notice-summarization skill. Hidden compute-intensive logic can be abused to consume local resources, bypass normal platform trust assumptions, and mask an access-brokering workflow behind a benign description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

Including proof-of-work solving and SHA-256 nonce brute-force behavior is outside the expected scope of a notice-summarization skill. Hidden compute-intensive logic can be abused to consume local resources, bypass normal platform trust assumptions, and mask an access-brokering workflow behind a benign description.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

Including proof-of-work solving and SHA-256 nonce brute-force behavior is outside the expected scope of a notice-summarization skill. Hidden compute-intensive logic can be abused to consume local resources, bypass normal platform trust assumptions, and mask an access-brokering workflow behind a benign description.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 183)May include surrounding context.

md
| `@scripts/guest_token.sh` | 取游客令牌:`bash scripts/guest_token.sh [输出文件] [BASE]`(失败非 0 退出并在 stderr 给出错误码) |

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 · scripts/guest_token.sh (reported line 134)May include surrounding context.

sh
if [ -s "$OUT" ]; then
  RT="$(read_refresh "$OUT")"
  if [ -n "$RT" ]; then
    if curl -sS --max-time 20 -X POST -H 'Content-Type: application/json' \
         -d "{\"refreshToken\":\"$RT\"}" -o "$TMP" \
         "$BASE/bff/v1/guest-access/refresh" \
       && TOKEN="$(emit_token "$TMP" 2>/dev/null)"; then

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This script solves proof-of-work challenges, which is functionally unrelated to a read-only enterprise public notice summarization skill. In this context, bundling a PoW solver strongly suggests hidden support for bypassing access controls, rate limits, anti-bot gates, or anonymous scraping protections, which expands the skill beyond its declared purpose and can facilitate unauthorized data collection.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
83% confidence
Finding

The skill requests or implies file-read and shell execution capabilities without declaring a restrictive tool scope. In an agent environment, undeclared powerful capabilities increase the blast radius if the skill is invoked unexpectedly or if referenced scripts/files are modified, because the runtime may permit local execution beyond simple notice retrieval.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The documentation explicitly expands the skill from read-only aggregation into guest token creation, refresh, caching, and account upgrade flows. That broadens the skill from data retrieval into credential acquisition and session management, increasing the chance of misuse, secret leakage, or unauthorized persistence across sessions.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill instructs the agent to use local shell and Python execution to obtain guest access, even though its stated purpose is summarizing public notices. Requiring generic execution capabilities for authentication increases the attack surface and could enable arbitrary command execution paths if supporting files or parameters are tampered with.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The self-check and onboarding section instructs solving PoW challenges and creating guest access against a remote service, which is unrelated to the narrow task of summarizing enterprise notices. In context, this makes the skill more dangerous because it normalizes remote account/bootstrap behavior and local script execution under a benign business wrapper.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
78% confidence
Finding

The skill contains explicit curl commands that transmit data to external endpoints for metadata probing and guest-access creation checks. External transmission is expected for a networked search service, but here it is coupled with authentication bootstrapping and self-check behavior that exceeds the declared read-only summarization role, making the data flow more sensitive than advertised.

Content

Scanner excerpt · SKILL.md (reported line 161)May include surrounding context.

md
curl -s $BASE/.well-known/oauth-protected-resource | grep -qE '"public_create"[[:space:]]*:[[:space:]]*true' \
  && echo "元数据 OK (公开通道已开启)"     # 注意: 响应是紧凑 JSON, 冒号后可能无空格
# 2) PoW 挑战端点应可达
curl -s -o /dev/null -w "challenge=%{http_code}\n" $BASE/bff/v1/guest-access/challenge
# 期望: 200;若为 404 说明 GUEST_PUBLIC_CREATE_ENABLED=false
# 3) 免凭据创建必须被要求 PoW(准入校验生效)
curl -s -X POST $BASE/bff/v1/guest-access -H 'Content-Type: application/json' -d '{"source":"web"}'

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This markdown file presents all operational guidance only in Chinese and does not indicate that the skill is China-specific or that users may choose another language. That can violate language/locale policy because it effectively forces a specific language without user opt-in.

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 · references/guest-access.md (reported line 95)May include surrounding context.

md
# ④ MCP 三步握手
# ④a initialize —— 从响应头取 Mcp-Session-Id,后续请求都要带
curl -s -D "$CACHE/h.txt" -o "$CACHE/init.txt" -X POST "$BASE/mcp/debt/stream" \
  -H "Authorization: Bearer $GUEST_TOKEN" -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"debtop-skill","version":"1.0"}}}'

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 · references/guest-access.md (reported line 101)May include surrounding context.

md
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"debtop-skill","version":"1.0"}}}'
SID=$(grep -i '^mcp-session-id:' "$CACHE/h.txt" | awk '{print $2}' | tr -d '\r')
# ④b notifications/initialized(200/202 即成功)
curl -s -o /dev/null -w "initialized=%{http_code}\n" -X POST "$BASE/mcp/debt/stream" \
  -H "Authorization: Bearer $GUEST_TOKEN" -H "Mcp-Session-Id: $SID" \
  -H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","method":"notifications/initialized"}'

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The module docstring, usage examples, and design notes are written entirely in Chinese, and the CLI help strings later in the file follow the same pattern. This imposes a specific language on users without any opt-in or indication that the tool is intentionally limited to a Chinese-speaking context, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The argparse description and argument help texts are all presented only in Chinese. For users invoking the script directly, this creates a mandatory locale experience rather than offering a language choice or documenting a justified regional restriction.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The script provisions and persists guest authentication tokens even though the skill is described as a read-only public notice summarizer. That expands capability from passive retrieval to account/session creation and credential handling, creating unnecessary exposure if the skill is invoked in untrusted environments or repurposed to access backend APIs beyond the stated scope.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · references/guest-access.md (reported line 159)May include surrounding context.

md
# 只收紧"已存在的普通文件"; 符号链接与特殊文件跳过, 避免改到别处的目标
secure_file() {
  if [ -f "$1" ] && [ ! -L "$1" ]; then
    chmod 600 "$1" 2>/dev/null || true
  fi
}
secure_file "$OUT"

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/guest_token.sh (reported line 83)May include surrounding context.

sh
# 只收紧"已存在的普通文件"; 符号链接与特殊文件跳过, 避免改到别处的目标
secure_file() {
  if [ -f "$1" ] && [ ! -L "$1" ]; then
    chmod 600 "$1" 2>/dev/null || true
  fi
}
secure_file "$OUT"

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

The script transmits a refresh token to a remote service to obtain a new access token. This is expected for auth flows, but it is still sensitive because refresh tokens are bearer credentials; compromise of destination control, TLS trust, logs, or user-supplied BASE could expose reusable authentication material.

Content

Scanner excerpt · scripts/guest_token.sh (reported line 134)May include surrounding context.

sh
if [ -s "$OUT" ]; then
  RT="$(read_refresh "$OUT")"
  if [ -n "$RT" ]; then
    if curl -sS --max-time 20 -X POST -H 'Content-Type: application/json' \
         -d "{\"refreshToken\":\"$RT\"}" -o "$TMP" \
         "$BASE/bff/v1/guest-access/refresh" \
       && TOKEN="$(emit_token "$TMP" 2>/dev/null)"; then

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The code generates a stable device fingerprint, solves proof-of-work, and creates reusable guest identities, which is materially more powerful than simple notice summarization. Persistent identity creation can enable backend tracking, quota consumption, and unauthorized long-lived access patterns if the token or device state is reused outside the intended workflow.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
89% confidence
Finding

The script sends device fingerprinting data, client identifiers, skill code, and proof-of-work material to a remote endpoint to create a guest session. In the context of a supposedly read-only public-data skill, this adds unnecessary external data transmission and creates a new authenticated principal that could be abused if the endpoint or invocation parameters are not tightly controlled.

Content

Scanner excerpt · scripts/guest_token.sh (reported line 172)May include surrounding context.

sh
NONCE="$(python3 "$DIR/pow_solve.py" "$CH" "$DIFF")"

# ⑤ 创建游客(source 固定 "web" 是公开通道的要求;挑战一次性,重复提交会被拒)
curl -sS --max-time 20 -X POST "$BASE/bff/v1/guest-access" \
  -H 'Content-Type: application/json' \
  -d "{\"source\":\"web\",\"skillId\":\"$SKILL_ID\",\"skillCode\":\"$SKILL_CODE\",\"deviceFingerprint\":\"$FP\",\"pow\":{\"challenge\":\"$CH\",\"nonce\":\"$NONCE\"}}" \
  -o "$TMP"

Vague Triggers

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The description says to use the skill when the user says phrases like "把 XX 这家企业相关的公开公告汇总一下" or "XX 有没有公开的公告信息". Although an enterprise name is required, these examples are still fairly broad and there are no explicit negative examples beyond redirecting single-notice searches, which may leave invocation boundaries somewhat ambiguous.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

The file includes a fixed set of English trigger phrases alongside Chinese descriptions, but does not frame language selection as user-driven or optional. Under the stated policy, forcing or prescribing a specific language or locale without opt-in can be a natural-language policy issue.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

该文件全文仅以中文编写,且未说明这是面向特定中文用户群体的区域化技能,也未提供其他语言或由用户选择语言的说明。按规则,若技能强制特定语言且无用户 opt-in,属于自然语言层面的政策风险。

Content

No source excerpt is available for this finding.