authority-runtime

v0.5.0 suspicious
8.0
High Risk

IAM + Context Control Plane for AI Agents — cryptographic policy enforcement with Ed25519 signed envelopes

🤖 AI Analysis

Final verdict: SUSPICIOUS

The package exhibits multiple high-risk behaviors including potential unauthorized data transmission, command injection vulnerabilities, and direct access to sensitive system files. These factors combined with low repository activity and lack of maintainer information suggest a high likelihood of malicious intent.

  • High shell risk due to execution of subprocess commands
  • Potential credential harvesting from accessing /etc/passwd
  • Low repository activity and unknown maintainer
Per-check LLM notes
  • Network: The network call to 'api/chat' is unusual and may indicate unauthorized data transmission.
  • Shell: Executing subprocess commands based on external input poses a high risk of command injection and unauthorized access.
  • Obfuscation: The use of base64 decoding on what appears to be a secret key suggests obfuscation or encryption of sensitive information, which could be legitimate but also suspicious without context.
  • Credentials: Direct access to system files like '/etc/passwd' indicates potential unauthorized credential harvesting, which is highly suspicious and likely malicious.
  • Metadata: The repository's low activity and the maintainer's lack of information raise concerns about potential malicious intent.

📦 Package Quality Overall: Medium (5.8/10)

✦ High Test Suite 9.0

Test suite present — 22 test file(s) found

  • Test runner config found: pyproject.toml
  • 22 test file(s) detected (e.g. test_archive.py)
◈ Medium Documentation 5.0

Some documentation present

  • Detailed PyPI description (7026 chars)
○ Low Contributing Guide 2.0

No contributing guide or governance files found

  • No CONTRIBUTING, CODE_OF_CONDUCT, or governance files found
◈ Medium Type Annotations 5.0

Partial type annotation coverage

  • 324 type-annotated function signatures detected in source
✦ High Multiple Contributors 8.0

Active multi-contributor project

  • 3 unique contributor(s) across 25 commits in tronmongoose/agent.carryall
  • Small but multi-author team (3–4 contributors)

🔬 Heuristic Checks

⚠ Outbound Network Calls score 3.0

Found 2 network call pattern(s)

  • api/chat" async with aiohttp.ClientSession() as session: async with session.post(
  • y: async with aiohttp.ClientSession() as session: async with session.post(ur
⚠ Code Obfuscation score 4.0

Found 2 obfuscation pattern(s)

  • "" secret_key_bytes = base64.b64decode(secret_key_b64) self.import_key(agent_id, secret_key
  • h) as f: loaded = __import__("json").load(f) assert loaded["report_type"] == "agent_acce
⚠ Shell / Subprocess Execution score 2.0

Found 1 shell execution pattern(s)

  • try: result = subprocess.run( self.mcp_command, input=jso
⚠ Credential Harvesting score 5.0

Found 2 credential access pattern(s)

  • ope secure_read(path="/etc/passwd") # Raises PermissionDenied # This works with val
  • ope secure_read(path="/etc/passwd", _envelope=valid_envelope) ``` """ def __
✓ Typosquatting

No typosquatting candidates detected

✓ Registered Email Domain

Email domain looks legitimate: authority-runtime.dev>

✓ Suspicious Page Links

All external links appear legitimate

⚠ Git Repository History score 2.5

Git history flags: Repository has zero stars and zero forks

  • Repository has zero stars and zero forks
⚠ Maintainer History score 4.0

2 maintainer concern(s) found

  • Author name is missing or very short
  • Author "" appears to have only 1 package on PyPI (new or inactive account)
✓ Known CVE Vulnerabilities

No known vulnerabilities found in OSV database.

💡 AI App Starter Prompt

Use this prompt to build a project with authority-runtime 
Develop a Python-based mini-application named 'SecureTaskManager' that leverages the 'authority-runtime' package to manage tasks securely within a distributed system environment. This application will allow users to create, assign, and execute tasks while ensuring that each task is executed only if it passes a strict security check enforced by cryptographic policies. Here’s a detailed breakdown of the project scope and requirements:

1. **Task Creation**: Users should be able to define new tasks, specifying the task details such as name, description, and required actions.
2. **Task Assignment**: Each task must be assigned to one or more agents, who are responsible for executing the task. The assignment process should include generating a cryptographic envelope using Ed25519 signatures that encloses the task details and specifies the necessary permissions for execution.
3. **Execution Authorization**: Before any agent executes a task, 'SecureTaskManager' must verify the authenticity and integrity of the task using the cryptographic envelope provided. Only if the verification succeeds, the task is allowed to proceed.
4. **Audit Logs**: Maintain a comprehensive audit log of all tasks including their creation, assignment, attempted execution, and actual execution status. This log should also record any failed verifications due to invalid cryptographic envelopes.

**Suggested Features**:
- Support for multiple users and roles within the system.
- Real-time notifications for task assignments and execution statuses.
- A user-friendly command-line interface (CLI) for interacting with 'SecureTaskManager'.
- Integration with existing logging frameworks for enhanced monitoring and debugging capabilities.

**Utilization of 'authority-runtime' Package**:
- Use 'authority-runtime' to generate cryptographic envelopes for each task, which includes the task details and the required permissions.
- Implement the verification logic using 'authority-runtime' to ensure that only tasks with valid cryptographic signatures are executed.
- Leverage the IAM capabilities of 'authority-runtime' to manage different roles and permissions across users and agents.

Your goal is to demonstrate a practical use case for the 'authority-runtime' package in securing task management within a distributed system, ensuring that every action taken is both secure and auditable.

💬 Discussion Feed

Leave a comment

No discussion yet. Be the first to share your thoughts!