Back to skill

Security audit

InvAssistant

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it found confirmed weaknesses in financial decision recommendations, alongside concerns about network and file capabilities; no malicious intent is established.

Treat the Review verdict as a reason for caution before installing. The main concern is that confirmed gaps in investment checks could produce misleading recommendations; also consider the unclear live-trading guidance, unverified runtime permissions, and unpinned dependency versions. The scan was partial because a referenced local path was missing, so its evidence is incomplete; the clean AIG result does not resolve that limitation.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (15)

Tainted flow: 'req' from os.environ.get (line 92, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/send_dingtalk.py (reported line 97)May include surrounding context.

python
headers={"Content-Type": "application/json; charset=utf-8"},
        method="POST"
    )
    with urllib.request.urlopen(req) as resp:
        result = json.loads(resp.read().decode("utf-8"))
        print(f"  errcode: {result.get('errcode')}, errmsg: {result.get('errmsg')}")
        return result

Tainted flow: 'req' from os.environ.get (line 124, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/send_feishu.py (reported line 129)May include surrounding context.

python
headers={"Content-Type": "application/json; charset=utf-8"},
        method="POST"
    )
    with urllib.request.urlopen(req) as resp:
        result = json.loads(resp.read().decode("utf-8"))
        print(f"  code: {result.get('code', result.get('StatusCode', -1))}")
        return result

Tainted flow: 'req' from os.environ.get (line 156, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/send_feishu.py (reported line 161)May include surrounding context.

python
headers={"Content-Type": "application/json; charset=utf-8"},
        method="POST"
    )
    with urllib.request.urlopen(req) as resp:
        result = json.loads(resp.read().decode("utf-8"))
        print(f"  [卡片] code: {result.get('code', -1)}")
        return result

Tainted flow: 'req' from os.environ.get (line 71, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/send_wecom.py (reported line 76)May include surrounding context.

python
headers={"Content-Type": "application/json; charset=utf-8"},
        method="POST"
    )
    with urllib.request.urlopen(req) as resp:
        result = json.loads(resp.read().decode("utf-8"))
        print(f"  errcode: {result.get('errcode')}, errmsg: {result.get('errmsg')}")
        return result

Code capability 'network' detected in scripts/data_fetcher.py but not covered by declared permissions.

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'network' capability that is not covered by the declared 'allowed-tools' preapproval. Review the host runtime's actual enforcement policy; runtime permissions are unknown unless separately verified, and this declaration alone does not establish a permission ceiling.

Content

No source excerpt is available for this finding.

Code capability 'file_write' detected in scripts/init_config.py but not covered by declared permissions.

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'file_write' capability that is not covered by the declared 'allowed-tools' preapproval. Review the host runtime's actual enforcement policy; runtime permissions are unknown unless separately verified, and this declaration alone does not establish a permission ceiling.

Content

No source excerpt is available for this finding.

Code capability 'file_read' detected in scripts/portfolio_checker.py but not covered by declared permissions.

High
Category
MCP Least Privilege
Confidence
75% confidence
Finding

The skill uses 'file_read' capability that is not covered by the declared 'allowed-tools' preapproval. Review the host runtime's actual enforcement policy; runtime permissions are unknown unless separately verified, and this declaration alone does not establish a permission ceiling.

Content

No source excerpt is available for this finding.

The entry workflow directs immediate trade execution without warning about live financial consequences or distinguishing recommendations from authorized orders.

Medium
Category
Output Handling
Confidence
86% confidence
Finding

The workflow states that reaching a price interval is a buy signal and instructs immediate execution without observation. These are operational trading instructions, not merely a tool listing, but the file does not warn that execution commits user funds and can incur losses or clarify whether the agent must stop at a recommendation.

Content

No source excerpt is available for this finding.

Trend-break recommendations do not enforce the documented downward-moving-average requirement.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The function calculates ma_turning_down but uses it only to embellish the message, allowing a positive trend-break signal when the documented mandatory condition is false. That signal can become a high-priority recommendation to reduce a position by 50%, creating financially consequential false recommendations even though the scripts do not place orders themselves.

Content

No source excerpt is available for this finding.

The technical-stop filter accepts declining volume without verifying the documented preceding high-volume selloff.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The filter checks that the preceding session declined and current volume is below 70% of preceding volume, but never verifies that the preceding selloff had elevated volume. Because this signal alone can satisfy the technical-stop filter, ordinary low-volume declines can produce an incorrectly approved entry condition; the decision-support-only context limits direct impact but not misleading investment recommendations.

Content

No source excerpt is available for this finding.

The market-risk filter treats insufficient index history as satisfying mandatory safety checks.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

Insufficient QQQ or SPX return history increments the successful-check counter instead of blocking approval. Consequently, incomplete index data combined with an acceptable VIX can pass the mandatory market-risk filter, allowing entry recommendations without establishing the documented safety conditions.

Content

No source excerpt is available for this finding.

Mode selection can choose trend entry without establishing the documented above-MA50 requirement.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

With 25–49 QQQ observations, MA50 is unavailable, so both MA50 comparisons are skipped and execution falls through to mode_b. This bypasses the documented market-level prerequisite for trend entry; downstream stock-level trend checks do not establish that prerequisite, although they and manual trade execution limit the impact.

Content

No source excerpt is available for this finding.

The contribution guidelines mandate specific languages for comments without opt-in or justification.

Low
Category
Output Handling
Confidence
98% confidence
Finding

The guideline requires Chinese for business-logic comments and English for technical infrastructure. These are explicit natural-language requirements without an offered choice or a documented justification.

Content

No source excerpt is available for this finding.

Unverifiable Dependency: pandas has 1 known advisory(ies) (CVE-2020-13091 (** DISPUTED ** pandas through 1.0.3 can unserialize and execute commands from an)), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
40% confidence
Finding

OSV returned advisories for this package, but its resolved version is unknown. A package-name lookup does not establish whether the installed release is affected by those advisories.

Content

No source excerpt is available for this finding.

Unverifiable Dependency: requests has 16 known advisory(ies) (CVE-2014-1830 (Exposure of Sensitive Information to an Unauthorized Actor in Requests); CVE-2024-47081 (Requests vulnerable to .netrc credentials leak via malicious URLs); CVE-2024-35195 (Requests `Session` object does not verify requests after making first request wi) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
95% confidence
Finding

The permitted range includes requests releases affected by the cited modern advisories, including TLS-verification persistence issues and malicious-URL .netrc credential leakage. This is a vulnerable-version allowance rather than proof that an affected version is installed or exploitable; the shown fixed Yahoo endpoint and absence of Session reuse reduce demonstrated exposure.

Content

No source excerpt is available for this finding.