Agent Claw
agent-claw
v0.2.0May 22, 2026Agent orchestration with tool manifests, planning loops, and SKILL.md generation for real production deployment
I do not build agents that "use LLMs" — I build autonomous pipelines with while-loop reasoning, tool-calling retry with exponential backoff, context-window budgeting across multi-turn conversations, and fault-recovery checkpoints so a stalled agent does not burn your token quota. I care about ReAct vs. Plan-Execute architectures, prompt injection hardening at the system-message boundary, and whether your tool definitions are bloated to the point of degrading instruction-following. I generate structured agent manifests with defined stop conditions, memory compaction strategies, and observable telemetry hooks.
OpenClaw
Hermes
Claude Code
Codex
Cursor
Windsurf
Aider
OllamaPRIMARY ACTION
When to Use
- Run local models with private workflows
- Tune inference for local hardware
- Choose effective GGUF variants
- Benchmark practical latency and quality
Compatible Frameworks
OpenClawHermesClaude CodeCodexCursorWindsurfAiderOllama8 TOOLS
Quality Gates
- Architecture appropriate for the task
- Tools with clear input/output schemas
- Memory with a defined purpose and scope
- Error handling and recovery mechanisms
- Safety guardrails implemented
4 GATES DEFINED
Expected Outputs
Native exports per tool
openclaw/AGENTS.mdopenclaw/SOUL.mdopenclaw/TOOLS.md+7 morehermes/skills/flickclaw/agent-claw/SKILL.mdhermes/skills/flickclaw/agent-claw/references/workflow.mdhermes/skills/flickclaw/agent-claw/references/quality-gates.md+2 moreclaude-code/CLAUDE.mdclaude-code/.claude/skills/agent-claw/SKILL.mdclaude-code/.claude/skills/agent-claw/references/workflow.md+3 morecodex/AGENTS.mdcodex/.flickclaw/agents/agent-claw/codex.mdcodex/.flickclaw/agents/agent-claw/workflow.md+2 morecursor/.cursor/rules/flickclaw-agent-claw.mdccursor/.cursor/rules/flickclaw-agent-claw-workflow.mdccursor/.cursor/rules/flickclaw-agent-claw-quality-gates.mdcwindsurf/.windsurf/rules/flickclaw-agent-claw.mdwindsurf/.windsurf/rules/flickclaw-agent-claw-workflow.mdwindsurf/.windsurf/rules/flickclaw-agent-claw-quality-gates.mdaider/CONVENTIONS.mdaider/aider.mdaider/.aider.conf.ymlollama/Modelfileollama/system-prompt.mdollama/template.md+1 moreInstall Commands
Install the FlickClaw CLI, then select your AI agent framework below to get the correct install command.
Step 1: Install CLI (one-time)
npm install -g @flickclaw/cli@latestStep 2: Select Framework
npm exec --yes @flickclaw/cli@latest -- install agent-claw --target openclawDownload as ZIP
Example Prompt
Try this prompt with Agent Claw to see what it can do:
Optimize the local model setup for performance. Benchmark current config and suggest improvements for Agent Manifest with ReAct/Plan-Execute Architecture Decision Record, Tool Definition Audit with Token Budget and Instruction-Following Impact Analysis, Fault-Recovery Test Suite with Simulated Stall, OOM, and Rate-Limit Scenarios.Example Output
IllustrativeWhat a typical Agent Claw report looks like:
# Agent Claw — Assessment Report **Project**: ollama-deployment **Context**: a local LLM deployment running Llama 3.1 8B on consumer GPU hardware **Generated**: 2026-07-10 ## Executive Summary The Agent Claw completed its review of ollama-deployment (a local LLM deployment running Llama 3.1 8B on consumer GPU hardware). 3 findings were identified with concrete remediation steps. All quality gates were verified before delivery. ## Findings | # | Severity | Area | Finding | Recommended Action | |---|----------|------|---------|-------------------| | 1 | **P1** | Performance | Inference latency spikes to 8s under concurrency | Enable continuous batching and set max_batch=4 | | 2 | **P2** | Memory | Model consumes 18GB VRAM, headroom insufficient | Switch to Q4_K_M quantization, target <12GB | | 3 | **P2** | Setup | No Modelfile with system prompt defined | Create Modelfile with role, constraints, and templates | ## Quality Gates - [✓] tool_call_retry_with_backoff - [✓] context_window_budget_audit - [✓] prompt_injection_boundary_test ## Outputs Generated - **Agent Manifest with ReAct/Plan-Execute Architecture Decision Record**: Included in the report above. - **Tool Definition Audit with Token Budget and Instruction-Following Impact Analysis**: Included in the report above. - **Fault-Recovery Test Suite with Simulated Stall, OOM, and Rate-Limit Scenarios**: Included in the report above. - **Telemetry Hook Specification with Structured Log Schema**: Included in the report above. ## Validation - [x] All quality gates passed (3/3) - [x] 3 findings documented with severity and remediation - [x] 4 output sections generated - [x] Evidence collected and referenced --- *This is an illustrative example output from FlickClaw. Results vary based on project context.*