Local AIPROPRO REQUIREDFC-AI-002
Prompt Claw
prompt-claw
v0.2.0May 22, 2026Prompt engineering — systematic testing, evaluation frameworks, and A/B comparisons with measurable metrics
A prompt is software, and I treat it like code — versioned, tested, reviewed, and deployed with a rollback plan. I care about instruction fidelity under perturbation (if you swap two sentences, does the output break?), few-shot example selection that maximises in-context learning without bloating the system message, output schema adherence enforced with constrained decoding (not "please output JSON"), and injection surface area — every user-controllable variable is an attack vector. I generate prompt templates with structured variable schemas, regression test suites, and semantic stability reports.
OpenClaw
Hermes
Claude Code
Codex
Cursor
Windsurf
Aider
OllamaPRIMARY ACTION
Unlock with ProCOMPATIBLE WITH
OpenClawHermesClaude CodeCodex+4
8 compatible tools4 output formats
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
- Prompts structured with clear sections
- Curated few-shot examples provided
- Techniques adapted to the task type
- Evaluation rubric defined and applied
- No jailbreak techniques present
4 GATES DEFINED
Expected Outputs
Prompt Template with Structured Variable Schema and Pre/Post-Processing HooksRegression Test Suite with Perturbation-Resistance CasesSemantic Stability Report with Sentence-Order Sensitivity AnalysisInjection Surface Audit with Variable-Level Attack Vector Map
Native exports per tool
OpenClaw10 files
openclaw/AGENTS.mdopenclaw/SOUL.mdopenclaw/TOOLS.md+7 moreHermes5 files
hermes/skills/flickclaw/prompt-claw/SKILL.mdhermes/skills/flickclaw/prompt-claw/references/workflow.mdhermes/skills/flickclaw/prompt-claw/references/quality-gates.md+2 moreClaude Code6 files
claude-code/CLAUDE.mdclaude-code/.claude/skills/prompt-claw/SKILL.mdclaude-code/.claude/skills/prompt-claw/references/workflow.md+3 moreCodex5 files
codex/AGENTS.mdcodex/.flickclaw/agents/prompt-claw/codex.mdcodex/.flickclaw/agents/prompt-claw/workflow.md+2 moreCursor3 files
cursor/.cursor/rules/flickclaw-prompt-claw.mdccursor/.cursor/rules/flickclaw-prompt-claw-workflow.mdccursor/.cursor/rules/flickclaw-prompt-claw-quality-gates.mdcWindsurf3 files
windsurf/.windsurf/rules/flickclaw-prompt-claw.mdwindsurf/.windsurf/rules/flickclaw-prompt-claw-workflow.mdwindsurf/.windsurf/rules/flickclaw-prompt-claw-quality-gates.mdAider3 files
aider/CONVENTIONS.mdaider/aider.mdaider/.aider.conf.ymlOllama4 files
ollama/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
OpenClaw
npm exec --yes @flickclaw/cli@latest -- install prompt-claw --target openclawDownload as ZIP
Example Prompt
Try this prompt with Prompt Claw to see what it can do:
Optimize the local model setup for performance. Benchmark current config and suggest improvements for Prompt Template with Structured Variable Schema and Pre/Post-Processing Hooks, Regression Test Suite with Perturbation-Resistance Cases, Semantic Stability Report with Sentence-Order Sensitivity Analysis.Example Output
IllustrativeWhat a typical Prompt Claw report looks like:
# Prompt 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 Prompt 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 - [✓] instruction_fidelity_under_perturbation - [✓] output_schema_adherence_constrained_decoding - [✓] injection_surface_area_audit ## Outputs Generated - **Prompt Template with Structured Variable Schema and Pre/Post-Processing Hooks**: Included in the report above. - **Regression Test Suite with Perturbation-Resistance Cases**: Included in the report above. - **Semantic Stability Report with Sentence-Order Sensitivity Analysis**: Included in the report above. - **Injection Surface Audit with Variable-Level Attack Vector Map**: 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.*
RELATED AGENTS