Back to skill

Security audit

Social By IDM

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent for managing social media, but it asks users to grant high-impact posting/deletion authority and includes credential and update practices that need careful review.

Install only if you are comfortable giving this skill authority over the selected social accounts. Use the narrowest API-key scopes possible, prefer the `X-Api-Key` header over `?key=...`, avoid sharing or screenshotting credential-bearing connector URLs, confirm before irreversible publish or delete actions, and avoid unpinned `npx` update commands unless you verify the package/version first.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:55
Finding

API Key Exposed Through URL Query Parameter

Content
View full analysis
` to the MCP URL and leave the OAuth fields blank. ``` ### Technical Analysis The Skill instructs users of URL-only clients to place the workspace API key in the MCP endpoint's query string. Query parameters are commonly retained by browser history, connector configuration, reverse-proxy and access logs, observability systems, diagnostic reports, screenshots, and support tooling. Unlike an authentication header, a query string forms part of the URL and is therefore more likely to be copied, displayed, or recorded by infrastructure handling the request. Although transport encryption protects the URL in transit from passive network observers, it does not prevent endpoints, proxies, or local tooling from recording it. ### Attack Path 1. A user follows the Skill instructions and configures an MCP URL containing `?key=`. 2. A client, reverse proxy, monitoring service, connector interface, or diagnostic facility records or displays the complete URL. 3. An attacker obtains access to the resulting log, configuration export, screenshot, browser history, or diagnostic report. 4. The attacker extracts the API key from the query string. 5. The attacker submits authenticated requests to the InstantDM MCP or REST service using the compromised key. 6. The attacker exercises every capability granted to that key until it is revoked. ### Impact Assessment The attacker's privileges are bounded by the scopes assigned to the exposed API key. Depending on those scopes, unauthorized actions could include: - Reading connected social-media account information. - Reading posts and publication status. - Upl ...[truncated 337 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:29
Finding

Unpinned npm Package Execution in Update Instructions

Content
View full analysis
Remediation
View remediation
update ``` 2. Document the expected npm package publisher, package URL, and trusted registry. 3. Publish and verify package integrity information or signed release provenance where supported. 4. Review the pinned release before changing the documented version. 5. Disable or avoid unexpected lifecycle scripts where practical. 6. Run update tooling in a restricted environment without unnecessary credentials or elevated privileges. 7. Prefer a verified local installation governed by a lockfile over resolving and executing a mutable package name on demand. ]]>
Vulnerability Patterns
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (4)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 100)May include surrounding context.

PUT (Content-Type: video/mp4, body: binary) POST /v1/media/{mediaId}/complete

text
List/delete: `GET /v1/media` · `DELETE /v1/media/{mediaId}`.

### Create post

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 124)May include surrounding context.

GET /v1/posts?status=scheduled&platform=instagram&limit=50 GET /v1/posts/{id} PUT /v1/posts/{id} (update caption, schedule, accounts, media — scheduled/draft only) DELETE /v1/posts/{id} (scheduled/draft only)

text

### Native platform posts (already-published)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 130)May include surrounding context.

Native platform posts (already-published)

text
GET    /v1/platform-posts?accountId=acc_xxx   → recent posts pulled from the platform
DELETE /v1/platform-posts/{id}                → returns { deleted: bool, reason }  (see gotchas)

Analytics

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.