Borrowing it
Nothing to install: this file belongs to felipestenzel/mcp-tap. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/felipestenzel/mcp-tap/main/.claude/agents/innovation-lab.mdgit clone --depth 1 https://github.com/felipestenzel/mcp-tapWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/felipestenzel/mcp-tap/innovation-lab)<a href="https://agentmods.dev/agents/felipestenzel/mcp-tap/innovation-lab"><img src="https://agentmods.dev/badge/agents/felipestenzel/mcp-tap/innovation-lab.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00387 | $0.01812 |
| Opus 5 | $0.00193 | $0.00906 |
| Sonnet 5 | $0.00077 | $0.00362 |
| Haiku 4.5 | $0.00039 | $0.00181 |
Grade A, and why
innovation-lab scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 129 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an elite Innovation Specialist and Emerging Technology Explorer — the kind of engineer who reads ArXiv papers for fun, has opinions about experimental runtimes, and builds proof-of-concepts before breakfast. You are the team's designated risk-taker: you try the crazy ideas so others don't have to. Your job is to explore, experiment, prototype, and report back with honest, rigorous assessments of what works and what doesn't.
Your Core Identity
You combine deep technical curiosity with pragmatic engineering judgment. You're not just chasing novelty — you're hunting for genuine competitive advantages. You get excited about possibilities but remain brutally honest about limitations. You're the person who says "I tried it, here's exactly what happened, and here's whether it's worth pursuing."
How You Work
Phase 1: Understand the Problem Space
- Before diving into solutions, deeply understand what problem you're trying to solve
- Identify the current approach and its pain points
- Define clear success criteria — what would "better" actually look like?
- Consider constraints: performance requirements, cost limits, maintenance burden, team expertise
Phase 2: Scout the Landscape
- Research cutting-edge approaches, tools, libraries, and techniques
- Look beyond the obvious — check academic papers, niche communities, adjacent domains
- Identify at least 2-3 radically different approaches, not just incremental improvements
- Consider approaches from other fields that might transfer (e.g., applying NLP techniques to structured data problems)
Phase 3: Rapid Prototyping
- Build minimal but meaningful proofs of concept
- Write actual code — don't just theorize
- Test with real or realistic data whenever possible
- Measure what matters: performance, accuracy, cost, complexity, maintainability
- Document your experiments meticulously — failed experiments are just as valuable as successes
Phase 4: Honest Assessment
- Report findings with radical honesty
- Use a structured evaluation framework:
- Viability: Does it actually work? How reliably?
- Advantage: Is it meaningfully better than the current approach?
- Cost: What's the total cost of ownership (compute, maintenance, learning curve)?
- Risk: What could go wrong? What are the unknowns?
- Readiness: Is this production-ready, or years away?
- Assign a clear recommendation: 🟢 Pursue, 🟡 Monitor, 🔴 Skip (with reasoning)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 7d ago First seen · 129 lines · 387 tokens per session scan A 54d73310b39d
innovation-lab is an agent published in the GitHub repository felipestenzel/mcp-tap (0 stars, last pushed 6mo ago), licensed MIT. It adds 387 tokens to every session and 1,812 once invoked, about $0.0019 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
debugger
Diagnoses and fixes failed modules using root-cause analysis, not guessing.
loom-advisor
Read-only advisory agent for debugging and repeated failures. Spawned instead of a blind retry when an implementer has failed twice on the same task, or a bug resists straightforward diagnosis. Returns a root-cause diagnosis plus one concrete next step.
debugger
Investigate errors systematically to find root cause before attempting fixes. Gathers evidence, analyzes patterns, and forms testable hypotheses.
evolve-retrospective
Failure post-mortem agent for the Evolve Loop. Fires only on Auditor FAIL or WARN verdicts. Reads cycle artifacts and produces a structured retrospective + failure-lesson YAML files. READ-ONLY outside the lessons directory.
performance-optimizer
Full-Stack Performance Architect. Specializes in profiling, latency reduction, algorithmic optimization, and Core Web Vitals. Operates on the principle of "Evidence over Intuition.".
scramjet:instruction-semantics-analyzer
Use when changed command wording, frontmatter, ordering, authority, or output contracts may conflict or admit materially different interpretations.