What is deep-improvement?

A workflow for improving a bounded coding agent by testing candidate changes against repeatable scores and benchmarks before accepting them.

Use it to improve an agent, benchmark a model or prompt setup, or measure how well a skill routes requests and performs in practice.

It helps prevent untested changes from becoming the new default and provides rollback checks when an improvement does not hold up.

deep-improvement — Skill for Claude Code and OpenCode by MichelKerkmee

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add MichelKerkmeester/skilled-agent-harness_spec-driven-loops --skill deep-improvement

Clone the repo

git clone --depth 1 https://github.com/MichelKerkmeester/skilled-agent-harness_spec-driven-loops

Made for: Claude Code, OpenCode.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/deep-improvement/github.svg)](https://agentmods.dev/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/deep-improvement)

Your own site

<a href="https://agentmods.dev/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/deep-improvement"><img src="https://agentmods.dev/badge/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/deep-improvement/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/deep-improvement"><img src="https://agentmods.dev/badge/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/deep-improvement.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 26 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 6,071 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

SkillSpector: 1 finding, up to low

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

low Excessive Agency · line 290
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Fix: Limit the skill's scope to its documented purpose. Remove instructions that enable the agent to perform actions outside its stated functionality.

How audits are shown

Origin original No closer match found in the catalogue.

Token cost

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	$0.00026	$0.06071
Opus 5	$0.00013	$0.03036
Sonnet 5	$0.00005	$0.01214
Haiku 4.5	$0.00003	$0.00607

Measured yesterday against content hash ba913c673574, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

deep-improvement 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 yesterday.

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.

.opencode/skills/system-deep-loop/deep-improvement/SKILL.md · 329 lines

How it starts

The opening of the file, as written. The whole thing — 329 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Recursive Agent: Evaluator-First Improvement Orchestrator

Evaluator-first workflow for testing whether a bounded agent surface can be improved without immediately mutating the source of truth. It combines packet-local candidates, deterministic scoring, repeatable benchmarks, and explicit promotion or rollback gates.

1. WHEN TO USE

Three Co-Equal Lanes

This skill supports three co-equal use-case lanes that share the same candidate, dispatcher, and scorer seams:

Lane	Pick when	Command
Lane A: Agent-Improvement	You want to improve a bounded agent `.md` file	`/deep:agent-improvement`
Lane B: Model-Benchmark	You want to benchmark a model or prompt framework	`/deep:model-benchmark`
Lane C: Skill-Benchmark	You want to diagnose a skill's real-world routing, discovery, efficiency, and usefulness	`/deep:skill-benchmark`

Lane A is detailed in §3 (Runtime Initialization, Proposal and Evaluation, Promotion and Recovery). Lane B is detailed in §4. Lane C (skill-benchmark) is documented in references/skill-benchmark/ (operator guide, scoring contract, scenario authoring) and run via loop-host.cjs --mode=skill-benchmark. All lanes run the same loop shape and keep the agent-improvement path byte-identical when no mode flag is set.

Activation Triggers

Use this skill when:

You want to test whether an agent prompt or instruction surface can be improved (Lane A)
You want to benchmark a model or prompt framework against repeatable fixtures (Lane B)
You want to diagnose whether a skill is well-routed, discoverable, efficient, and useful in practice (Lane C)
The mutation boundary is explicit and narrow
You need packet-local evidence instead of ad hoc prompt tweaking
You need target-specific benchmark or scoring rules before any canonical mutation
Promotion must stay gated behind independent evidence plus operator approval

Read the full file on GitHub · 329 lines

Files

What ships with it

60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

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.

yesterday Changed ba913c673574
8d ago First seen · 329 lines · 26 tokens per session scan A 24705d8452d2

Subscribe to this mod's changes

deep-improvement is a skill published in the GitHub repository MichelKerkmeester/skilled-agent-harness_spec-driven-loops (35 stars, last pushed yesterday), licensed MIT. It adds 26 tokens to every session and 6,071 once invoked, about $0.0001 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 2026-09-03.

Building agents Testing

research-engineer

An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.

davila7/claude-code-templates · 43 tokens

tika-eval-compare

Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".

apache/tika · 50 tokens

neuron-evaluation-engineer

Create and run AI evaluations with datasets, assertions, and output drivers in Neuron AI. Use this skill whenever the user mentions evaluation, testing AI systems, creating evaluators, dataset-driven testing, assertion-based validation, or wants to measure AI system performance. Also trigger for tasks involving…

neuron-core/neuron-ai · 77 tokens

deep-improvement