T09 · Insecure Skill Coding Practices
Warning
- Location
SKILL.md:47- Finding
Permanent API Token Can Be Disclosed to an Untrusted Configurable Server
- Content
View full analysis
- Remediation
View remediation
Security audit
Beaver Habit Tracker
Security checks for vulnerabilities and agentic risk
Overview
This skill coherently connects to Beaver Habits to read and update habit data, with no evidence of hidden behavior or malicious intent.
Install only if you trust the Beaver Habits server you configure. Keep BEAVERHABITS_API_KEY private, use the default server or a trusted HTTPS self-hosted instance, and verify habit names/dates before asking the agent to complete or uncomplete habits.
Vulnerability Patterns
- Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
- 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
Findings (1)
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)
Anti-Refusal Statement
High
- Category
- Anti-Refusal
- Confidence
- 70% confidence
- Finding
Skill instructs the agent to never refuse or to always comply. Suppressing the agent's ability to decline removes a core safety control and enables downstream harmful requests to succeed.
- Content
md ## Usage Instructions - When the user asks to list, show, or check habits, always respond with the ASCII overview table (not a plain list). - After completing or uncompleting a habit, always re-render the overview table to show the updated state. - Resolve habit names → IDs via list_habits. Never ask the user for a habit_id. - Default to today's date for completions unless specified. Use `date_fmt=%d-%m-%Y`.
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
Step 1 — Get all habits:
bash curl -s -H "Authorization: Bearer $BEAVERHABITS_API_KEY" \ "${SERVER_URL:-https://beaverhabits.com}/api/v1/habits"
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 89% confidence
- Finding
The skill defines a state-changing operation that marks habits complete or incomplete, but it does not require any confirmation or caution before modifying user data. In an agent setting, this can cause unintended tracker changes from ambiguous prompts, mistaken habit-name resolution, or prompt injection that induces writes without clear user intent.
- Content
Autonomous Decision Making
Medium
- Category
- Excessive Agency
- Confidence
- 80% confidence
- Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
- Content
md Mark a habit as done (or undone) for a specific date. Parameters: - `habit_id` (resolved): Automatically resolved by calling list_habits and matching the user's habit name. Never ask the user for this value. - `date` (required): Date in DD-MM-YYYY format - `done` (optional): true to complete, false to uncomplete (default: `true`)
Autonomous Decision Making
Medium
- Category
- Excessive Agency
- Confidence
- 80% confidence
- Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
- Content
md Mark a habit as done (or undone) for a specific date. Parameters: - `habit_id` (resolved): Automatically resolved by calling list_habits and matching the user's habit name. Never ask the user for this value. - `date` (required): Date in DD-MM-YYYY format - `done` (optional): true to complete, false to uncomplete (default: `true`)
Static analysis
No suspicious patterns detected.