Back to skill

Security audit

openfinance

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent OpenFinance helper skill that accesses sensitive financial data as its stated purpose, with no hidden execution or persistence found.

Install only if you intend the agent to access OpenFinance-connected account balances and transaction history. Keep OPENFINANCE_API_KEY private, request only needed fields or date ranges, and avoid posting reconnect links or financial responses in shared chats or logs.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (3)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs users to query bank accounts and transactions through remote API endpoints without any explicit warning that this is highly sensitive financial data. In an AI-agent context, users may not realize the privacy implications of transmitting balances, account metadata, and transaction histories to an external service, increasing the risk of unintended disclosure.

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 · SKILL.md (reported line 119)May include surrounding context.

Approved functions: COUNT, SUM, AVG, MIN, MAX, ABS, ROUND, CEIL, CEILING, FLOOR, LOWER, UPPER, LENGTH, TRIM, BTRIM, LTRIM, RTRIM, SUBSTRING, SUBSTR, REPLACE, DATE_TRUNC, DATE_PART, EXTRACT, TO_CHAR. CASE, COALESCE, NULLIF, GREATEST, LEAST, CURRENT_DATE, and CURRENT_TIMESTAMP are also supported. Casts are limited to built-in boolean, integer, numeric, floating-point, text, date, timestamp, and interval types. Arithmetic, comparisons, LIKE/ILIKE, IN, BETWEEN, and NULL/boolean tests are supported; custom operators, regex operators, COLLATE, and unlisted syntax are rejected. Function and type names may be unqualified or qualified with pg_catalog.

bash
curl -s -X POST "$BASE_URL/api/transactions/query" \
  -H "$AUTH_HEADER" \
  -H "Content-Type: application/json" \
  -d '{"sql": "SELECT SUM(amount), COUNT(*) FROM txns WHERE merchant_name ILIKE '\''%starbucks%'\''"}' \

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The reconnect-link feature creates a bearer-style public URL that can facilitate bank account reconnection, but the documentation does not emphasize that this link is sensitive and should only be shared with the account owner through a secure channel. If exposed in logs, chats, or to the wrong recipient, an unauthorized party may be able to initiate or interfere with the reconnection flow and gain sensitive institution context.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.