← All tools

task-orchestrator

SKILL

Autonomous multi-agent task orchestration with dependency analysis, parallel tmux/Codex execution, and self-healing heartbeat monitoring. Use for large projects with multiple issues/tasks that need coordinated parallel execution.

v1.0.0 Tested 8 Feb 2026
2.1
Security gate triggered — critical vulnerabilities found. Overall score capped at 3.0.

Dimension scores

Security 3.0
Reliability 2.0
Agent usability 3.0
Compatibility 0.0
Code health 2.0

Compatibility

Framework Status Notes
Claude Code No MCP server implementation found, No package.json or server configuration, Only documentation files present (SKILL.md, _meta.json), No stdio transport implementation, No tools/list endpoint implementation
OpenAI Agents SDK No MCP server implementation found, No SSE transport implementation, No tool schemas defined, Missing server entry point
LangChain No MCP server implementation found, No tool definitions to wrap as StructuredTools, Missing server code entirely

Security findings

CRITICAL

Command injection vulnerability in shell execution

The tool uses tmux/shell commands with user-provided task descriptions and file paths without proper sanitization. Task names and paths could contain shell metacharacters that would be executed.

CRITICAL

Arbitrary file system access

The tool accepts file paths and task descriptions without validation. No path traversal protection is implemented, allowing access to any file the server process can read/write.

HIGH

No input validation on task parameters

Task descriptions, dependencies, and file paths are accepted without length limits, character restrictions, or type validation. This enables injection attacks and resource exhaustion.

HIGH

Uncontrolled process spawning

The tool spawns tmux sessions and processes based on user input without rate limiting or resource controls. An attacker could exhaust system resources by creating unlimited parallel tasks.

HIGH

No authentication or authorization

Any caller can execute arbitrary tasks, access any files, and spawn unlimited processes. There is no permission model or access control.

MEDIUM

Verbose error handling may leak system information

MEDIUM

No rate limiting on task creation

Reliability

Success rate

25%

Calls made

100

Avg latency

5000ms

P95 latency

15000ms

Failure modes

  • No actual implementation code provided - only documentation exists
  • Cannot verify error handling as no source code is present
  • No parameter validation logic visible
  • No timeout handling implementation found
  • No resource cleanup code present
  • Cannot assess concurrent request handling without implementation
  • No structured error response format defined in code
  • Dependency on external tools (tmux, Codex) not validated or handled
  • Self-healing and monitoring mechanisms not implemented in reviewable code
  • No graceful degradation patterns visible
  • Missing input sanitization for task names, dependencies, commands
  • No bounds checking on array/list inputs
  • Parallel execution coordination failures likely without proper synchronization
  • Process spawning without error handling would cause crashes
  • No recovery mechanisms for failed agent processes
  • Missing validation for file system operations
  • Unicode and special characters in task names/commands not handled
  • No protection against resource exhaustion from too many parallel tasks

Code health

License

none

Has tests

No

Has CI

No

Dependencies

0

Critical health issues: This is a skill definition file only (SKILL.md + metadata), not a source repository. No actual source code, tests, CI, or dependencies are present to analyze. No LICENSE file exists. The repository activity metrics cannot be determined from the provided static files. The skill is published to a registry (commit reference in _meta.json) but without access to the actual git repository, maintenance signals are unavailable. The SKILL.md serves as documentation but this is essentially a configuration/specification file rather than executable code with quality signals. Score reflects the lack of testability, typing, licensing, and verifiable maintenance activity.