Skip to content

LeoYeAI/agent-master-skills / skills/agent-harness

agent-harness

Verified

Use when an agent should drive a goal in one of this repo's domains to a verified close, before starting the work. Compiles the goal into a task plan with goal_compiler.py and runs it through loop_controller.py, which verifies each task, caps retries and escalates.

assets/harness_manifest.schema.json

View on GitHub
{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "agent-harness domain manifest (agent-harness/manifest.v1)",
  "description": "Machine-readable inventory of one domain folder: every skill, its tools, its verification checks, and its agentic signals. Produced by harness_manifest_builder.py; consumed by goal_compiler.py. Regenerate after any skill change — never hand-edit.",
  "type": "object",
  "required": ["schema", "domain", "skill_count", "loop_defaults", "skills"],
  "properties": {
    "schema": { "const": "agent-harness/manifest.v1" },
    "domain": { "type": "string", "description": "Domain folder relative to repo root, e.g. 'engineering-team'." },
    "skill_count": { "type": "integer", "minimum": 0 },
    "generated_at": { "type": "string", "description": "UTC ISO-8601; omitted when built with --no-timestamp for diff-stable commits." },
    "loop_defaults": {
      "type": "object",
      "required": ["max_attempts_per_task", "max_loop_iterations", "escalate_on"],
      "properties": {
        "max_attempts_per_task": { "type": "integer", "minimum": 1 },
        "max_loop_iterations": { "type": "integer", "minimum": 1 },
        "escalate_on": { "type": "array", "items": { "type": "string" } }
      }
    },
    "skills": {
      "type": "array",
      "items": {
        "type": "object",
        "required": ["name", "path", "description", "tools", "agentic_signals"],
        "properties": {
          "name": { "type": "string" },
          "path": { "type": "string", "description": "Skill dir (contains SKILL.md) relative to repo root." },
          "description": { "type": "string", "maxLength": 600 },
          "tools": {
            "type": "array",
            "items": {
              "type": "object",
              "required": ["script", "wired", "supports_sample", "verification"],
              "properties": {
                "script": { "type": "string" },
                "wired": { "type": "boolean", "description": "True if the script basename appears in the skill's SKILL.md (A3 wiring)." },
                "supports_sample": { "type": "boolean" },
                "verification": {
                  "type": "array",
                  "items": {
                    "type": "object",
                    "required": ["cmd", "expect_exit", "kind"],
                    "properties": {
                      "cmd": { "type": "string" },
                      "expect_exit": { "type": "integer" },
                      "kind": { "enum": ["smoke", "sample", "manual-evidence"] }
                    }
                  }
                }
              }
            }
          },
          "agentic_signals": {
            "type": "object",
            "description": "Static evidence of agentic structure in SKILL.md (maps to audit dimensions AR1/AR4/AR5/AR6).",
            "required": ["goal_intake", "refusal_gate", "verification", "loop_discipline", "close_out"],
            "additionalProperties": { "type": "boolean" }
          },
          "references": { "type": "array", "items": { "type": "string" } }
        }
      }
    }
  }
}
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexUnverifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall