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.
npx skills add rajkaria/boil-the-ocean --skill boil-oceangit clone --depth 1 https://github.com/rajkaria/boil-the-oceanWrote 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/skills/rajkaria/boil-the-ocean/boil-ocean)<a href="https://agentmods.dev/skills/rajkaria/boil-the-ocean/boil-ocean"><img src="https://agentmods.dev/badge/skills/rajkaria/boil-the-ocean/boil-ocean/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.
<a href="https://agentmods.dev/skills/rajkaria/boil-the-ocean/boil-ocean"><img src="https://agentmods.dev/badge/skills/rajkaria/boil-the-ocean/boil-ocean.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00094 | $0.01935 |
| Opus 5 | $0.00047 | $0.00967 |
| Sonnet 5 | $0.00019 | $0.00387 |
| Haiku 4.5 | $0.00009 | $0.00194 |
Grade A, and why
boil-ocean 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 11d 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 — 151 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Boil the Ocean
Overview
Execute an entire multi-sprint spec to completion in one autonomous run. The user has
pre-authorized every decision below the only-stop line; the deliverable is the finished,
verified product — not a plan, not sprint 1 plus options. All run memory lives in
.ocean/ files, so any fresh session (daemon-relaunched or human-started) continues
with zero conversation history.
This file is the binding protocol for EVERY agent working an ocean run — on Claude Code it loads as a skill; on Codex, Gemini CLI, or any other agent it arrives as "read this file and follow it exactly". The rules are identical either way.
Scoping down is failure. When the user asked for everything, proposing "let's start with X and see" violates the contract. The scheduler gives you unlimited sessions — session limits are a relaunch, not a reason to shrink scope.
Violating the letter of this skill is violating its spirit. There is no "I followed the idea but skipped the checkpoint" — the checkpoint IS the idea.
The contract
| The user grants | You guarantee |
|---|---|
| Full decision authority below the only-stop line | Every decision logged in .ocean/DECISIONS.md |
| Multi-session runtime (daemon relaunches you) | Checkpointed state before every session end |
| No mid-run reviews | One commit per sprint, tests green before "done" |
| Trust | A final REPORT.md flagging anything that needs human review |
State files — the run's only memory
| File | Purpose | Rule |
|---|---|---|
.ocean/state.json |
Machine state: status, sprints, checkpoint | Mutate ONLY via ocean.sh — never edit by hand |
.ocean/PLAN.md |
Full sprint plan with acceptance criteria | Written once in planning, amended only with a DECISIONS entry |
.ocean/DECISIONS.md |
Decision journal | Append-only |
.ocean/SPEC.md |
The spec, if it arrived as pasted text | Immutable |
.ocean/REPORT.md |
Final report | Written at completion |
ocean.sh lives at scripts/ocean.sh in the boil-the-ocean install (when running as
a daemon worker, the launch prompt gives you its absolute path; if installed via
install.sh it is also on PATH as ocean).
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.
- 11d ago First seen · 151 lines · 94 tokens per session scan A 805fec84a044
boil-ocean is a skill published in the GitHub repository rajkaria/boil-the-ocean (2 stars, last pushed 2mo ago), licensed MIT. It adds 94 tokens to every session and 1,935 once invoked, about $0.0005 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 skills, from other repositories
sdlc-accelerate
End-to-end SDLC ramp-up from idea to construction-ready with automated phase transitions.
config-validator
Validate AIWG configuration files and project setup for correctness and completeness.
summarize-transcript
Analyze and summarize a transcript, meeting notes, or discussion thread into a clear actionable document.
aiwg-regenerate-copilot
Regenerate copilot-instructions.md for GitHub Copilot with vendor-specific content only.
artifact-metadata
Manage artifact metadata, versioning, ownership, and review history across the SDLC lifecycle.
hook-regenerate
Rebuild AIWG hook files from currently installed framework manifests.