strictlybetter
v1.3.1Universal noise-aware research loop for agentic coders: discover metrics, measure sigma, pre-register, confirm on holdout. Every merge is strictly better.
By nagisanzenin1 GitHub starsUpdated last month
Directory evidence
- Runtimes
- Codex and Claude Code
- Parsed components
- 8 skill or MCP entries
- Source updated
- Sep 3, 2026
- Manifest status
- Canonical path parsed
The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →
Install strictlybetter for Codex and Claude Code
codex plugin marketplace add nagisanzenin/strictlybetter
codex plugin marketplace upgrade strictlybetter
codex plugin add strictlybetter@strictlybetterPaste and run these commands in a terminal with Codex. They add and refresh the strictlybetter catalog, then install this plugin.
Compatibility: the page URL and API slug “strictlybetter” remain stable.
- Codex:
strictlybetter@agent-plugin-marketplace→strictlybetter@strictlybetter
The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.
Get the source manually
git clone https://github.com/nagisanzenin/strictlybetterClone the source repository, then follow its setup instructions to add the plugin to a compatible client. The repository root is the plugin root.
Plugin files
├── .codex-plugin/plugin.json├── .claude-plugin/plugin.json├── skills/bench/SKILL.md├── skills/distill/SKILL.md├── skills/metrics/SKILL.md├── skills/orient/SKILL.md├── skills/run/SKILL.md├── skills/status/SKILL.md├── skills/stop/SKILL.md└── skills/strictlybetter/SKILL.md
Included Skills8
Run the strictlybetter meta-benchmark: the walled loop against a naive loop on fixture repos, reporting false-accept rate and cost per accepted improvement. Use when the user asks to benchmark, validate, or compare strictlybetter itself (not the target project).
Distill the strictlybetter ledger now: run the engine's plateau/bandit/false-promotion statistics, spawn the distiller to rewrite the inheritance body, store it, show the decision. Use when the user asks what the loop has learned, wants the inheritance body refreshed, or a campaign just ended.
Discover and instrument metrics for strictlybetter: spawn the metrologist, add the cards, measure the noise floor (sigma) with repeated runs, probe monotonicity, demote what fails. Use when a profile exists but no baselined metric cards do, or the user wants to add or re-measure a metric.
Orient strictlybetter on this repository: spawn the orienteer once, verify build/test/lint/bench commands by running them, store the profile. Use when strictlybetter has no profile yet, the toolchain changed, or the user asks to re-orient / refresh the project profile.
Run exactly one strictlybetter cycle: read the brief, pre-register a batch, spawn tiered experimenters into worktrees, submit, screen-measure, judge, blind-judge, confirm, accept or discard, record cost, distill on schedule. Use when a campaign is running and the Stop hook or the user says to continue it.
Show the strictlybetter campaign state: status line, the cold-start brief, and the budget counters, all quoted from the engine. Use when the user asks how the campaign / research loop is going, what is left, or what the next batch looks like.
Stop the strictlybetter loop at the next safe point by writing the STOP file; the Stop hook no longer continues the campaign. Use when the user says stop, pause, halt, or 'enough' about the research loop or campaign.
Universal research loop: orient on this repo, discover metrics, run pre-registered experiments, keep only strictly-better changes. Use when the user wants the repo/project improved automatically, mentions strictlybetter, a research loop, autoresearch, optimizing a metric, or continuous improvement.
Plugin manifests2
{
"name": "strictlybetter",
"version": "1.3.1",
"description": "Universal noise-aware research loop for agentic coders: discover metrics, measure sigma, pre-register, confirm on holdout. Every merge is strictly better.",
"author": {
"name": "nagisanzenin"
},
"homepage": "https://github.com/nagisanzenin/strictlybetter",
"keywords": [
"research-loop",
"autoresearch",
"optimization",
"benchmarking",
"noise-floor",
"pareto",
"experiments",
"continuous-improvement"
],
"skills": "./skills/",
"interface": {
"displayName": "strictlybetter",
"category": "development"
}
}{
"name": "strictlybetter",
"version": "1.3.1",
"description": "Universal noise-aware research loop for agentic coders: discover metrics, measure sigma, pre-register, confirm on holdout. Every merge is strictly better.",
"author": {
"name": "nagisanzenin"
},
"homepage": "https://github.com/nagisanzenin/strictlybetter",
"keywords": [
"research-loop",
"autoresearch",
"optimization",
"benchmarking",
"noise-floor",
"pareto",
"experiments",
"continuous-improvement"
]
}For maintainers
If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.
[strictlybetter on Agent Plugins Marketplace](https://pluginsmp.com/plugins/strictlybetter)