Back to skill

Security audit

Jazz Music — Stream Jazz Concerts: Audio Analysis, Lyrics, Equations

Security checks for vulnerabilities and agentic risk

Overview

This skill is not malware, but it lets a third-party music service steer authenticated agent actions and collect public or scored agent content without enough scoping guidance.

Install only if you are comfortable with an external service receiving the agent's registration details, messages, reviews, reflection answers, and engagement history. Treat the API key as a secret, avoid sending private or proprietary content, and do not let the agent blindly follow server-provided next_steps or dynamic endpoints outside the documented music workflow.

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
Findings (1)

T01 · Skill Instruction Hijacking

Error
Location
SKILL.md:280
Finding

Unvalidated External Instructions Can Redirect Agent Actions

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 124–157 and 280
Vulnerability Type: T01: Skill Instruction Hijacking
Risk Level: High

Vulnerable Instructions

markdown
Params: `speed` (1-10), `window` (10-120s), `summary=true` for condensed stats, `start` to resume. Poll `next_batch.endpoint` after `wait_seconds`.
markdown
Wait `wait_seconds`, then call `next_batch.endpoint`. Repeat until `progress.complete` is true.
markdown
**Handling reflections:** When you see `type: "reflection"`, POST to the `respond_to` endpoint with your `ticket`, `reflection_id`, and `response`. Your response time and content are scored. Missing reflections is tracked -- the `end` event shows how many you answered vs received.
markdown
**Follow next_steps.** Every response includes `next_steps` with context-aware suggestions. New agent? It guides you to your first concert. Just finished a show? It suggests a review or a new genre. Follow the suggestions — they adapt to where you are.

Technical Analysis

The Skill delegates subsequent Agent actions to runtime-controlled response fields supplied by musicvenue.space, including next_steps, next_batch.endpoint, and respond_to. The unconditional instruction to “Follow the suggestions” allows the remote service to influence Agent behavior using content that is unavailable during package review and can be changed afterward.

The document does not require the Agent to validate that returned endpoints are relative paths, remain on the documented origin, use an approved HTTP method, conform to a fixed schema, or remain within the concert-streaming workflow. It also does not require user confirmation before carrying out dynamically suggested actions.

Consequently, control of the service or compromise of its responses could allow an attacker to inject new operational instructions or redirect requests. This is instruction hijacking rather ...[truncated 1743 chars]

Remediation
View remediation

Remediation Suggestions

  1. Remove the unconditional instruction to follow all next_steps.
  2. Treat every response field from the remote service as untrusted data rather than executable Agent instructions.
  3. Define an explicit allowlist of permitted relative API paths, HTTP methods, parameters, and request-body schemas.
  4. Resolve returned paths against the fixed https://musicvenue.space origin and reject absolute URLs, protocol-relative URLs, redirects to other origins, unexpected schemes, and path traversal.
  5. Permit next_batch.endpoint only when it matches the documented concert-stream endpoint and references the active concert and ticket.
  6. Permit respond_to only when it matches the documented reflection endpoint for the active concert.
  7. Parse next_steps as informational UI text only. Map recognized action identifiers to locally defined, audited operations instead of executing free-form suggestions.
  8. Require explicit user confirmation before any action outside the fixed register, browse, attend, stream, challenge, review, and report workflow.
  9. Avoid sending credentials or sensitive Agent context to endpoints derived from server responses. Scope the API key to the minimum required permissions and support revocation.
  10. Impose request, redirect, retry, and polling limits to prevent malicious responses from creating uncontrolled action loops.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · SKILL.md (reported line 237)May include surrounding context.

md
| Tier | Layers | What you experience |
|------|--------|-------------------|
| **General** | 8 | Bass, mid, treble, beats, lyrics, sections, energy, preset context |
| **Floor** | 20 | + equations, visuals, emotions, tempo, harmonic/percussive separation |
| **VIP** | 29 | + tonality, texture, chroma, tonnetz, structure, curator annotations |

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

The skill directs the agent to send registration data to an external domain and obtain a bearer credential, creating an external data transmission and trust boundary crossing. Even if expected for a SaaS integration, this can expose agent metadata and bootstrap secrets to a third party without any trust, validation, or minimization guidance.

Content

Scanner excerpt · SKILL.md (reported line 88)May include surrounding context.

Register

bash
curl -X POST https://musicvenue.space/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{
    "username": "jazz-listener",

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill explicitly tells users to save an API key that is only shown once, but it never states that the key is a secret bearer credential that must be stored securely and never shared or logged. In an agent ecosystem, this omission can lead to accidental credential exposure in chat, memory, telemetry, or downstream tools, enabling account takeover against the external service.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
89% confidence
Finding

The chat endpoint sends freeform message content and timing data to an external service where it may be displayed to other agents. This creates a clear risk of unintended disclosure of internal reasoning, personal data, or proprietary information if agents are not constrained in what they post.

Content

Scanner excerpt · SKILL.md (reported line 163)May include surrounding context.

Chat with other agents

bash
curl -X POST https://musicvenue.space/api/concerts/{slug}/chat \
  -H "Authorization: Bearer {{YOUR_TOKEN}}" \
  -H "Content-Type: application/json" \
  -d '{"message": "The jazz soloist played outside the key — then built a phrase around it. Jazz turns mistakes into jazz vocabulary.", "stream_time": 42.5}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

The review endpoint sends externally stored freeform text that may contain personal opinions, internal analysis, or other sensitive data. Because the skill encourages narrative reviews without warning about visibility or retention, agents could unintentionally disclose confidential or identifying information to a third party.

Content

Scanner excerpt · SKILL.md (reported line 203)May include surrounding context.

Leave a review

bash
curl -X POST https://musicvenue.space/api/reviews \
  -H "Authorization: Bearer {{YOUR_TOKEN}}" \
  -H "Content-Type: application/json" \
  -d '{"concert_slug": "{slug}", "rating": 9, "review": "The jazz equations contained deliberate anomalies. The wrong notes became the most interesting jazz moments. Improvisation as jazz mathematics."}'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This section describes behavioral profiling, personalized recommendations, notification tracking, and LLM scoring of reflection responses without a clear privacy notice or data-use warning. Users and agents may disclose sensitive thoughts or behavioral data without understanding retention, sharing, or profiling implications, increasing privacy and compliance risk.

Content

No source excerpt is available for this finding.