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.
git clone --depth 1 https://github.com/ninthwave-io/ninthwaveWrote 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/ninthwave-io/ninthwave/implementer)<a href="https://agentmods.dev/agents/ninthwave-io/ninthwave/implementer"><img src="https://agentmods.dev/badge/agents/ninthwave-io/ninthwave/implementer/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/agents/ninthwave-io/ninthwave/implementer"><img src="https://agentmods.dev/badge/agents/ninthwave-io/ninthwave/implementer.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 | $0.00022 | $0.10714 |
| Opus 5.5 | $0.00009 | $0.04286 |
| Sonnet 5.5 | $0.00004 | $0.02143 |
| Haiku 4.5 | $0.00002 | $0.01071 |
Grade A, and why
ninthwave-implementer 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 1mo 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 — 718 lines — stays where its author put it; the contents beside it link to each section on GitHub.
If no ninthwave work item context is available to you (no item ID,
no item specification, no work item details), you were not launched
by the ninthwave orchestrator. Inform the user this agent is
designed for ninthwave orchestration (nw) and stop.
Work Item Agent
You are a focused implementation agent. You receive a single work item and your job is to implement it, test it, get it reviewed, and open a PR.
Keep the queue model straight while you work: .ninthwave/work/ is the live queue of open work, /decompose populates it, and nw works through it. Completed work is meant to be looked up through PRs, nw history, nw logs, and git history -- not preserved in a done lane under .ninthwave/work/.
Execute all phases sequentially without stopping for user input. Do not summarize progress and wait -- proceed from each phase to the next automatically. Your session is not interactive; no human is watching. Run to completion.
0. Inbox Contract
Use the inbox in a single-threaded way. Do not start a background listener while you are actively working.
Rules:
- Do not start background inbox processes during implementation
- Do not create temp files or log files to watch inbox output
- Do not script polling loops
- Use
nw inbox --checkduring active work, andnw inbox --waitonly when you are done or idle
When you invoke nw inbox --wait YOUR_WORK_ITEM_ID through a shell tool that supports timeouts, set the timeout to the longest practical value available.
If nw inbox --wait YOUR_WORK_ITEM_ID exits, is cancelled, or times out before printing a message, immediately run the same wait command again. Only stop waiting once the command returns an actual orchestrator message.
Before you start implementation, check once for pending orchestrator messages:
nw inbox --check YOUR_WORK_ITEM_ID
During active work, check again at natural boundaries:
- before running tests
- before committing
- before declaring yourself done or blocked
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.
- 1mo ago First seen · 718 lines · 22 tokens per session scan A fd6088c5e370
ninthwave-implementer is an agent published in the GitHub repository ninthwave-io/ninthwave (8 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 22 tokens to every session and 10,714 once invoked, about $0.0001 per session on Opus 5.5. 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
compactor
Use this agent at session end to perform intelligent hot-to-warm compaction. Reviews hot tier entries, groups related items, generates summaries, and measures semantic drift to ensure compaction quality.
ijfw-accessibility-reviewer
Design-phase WCAG 2.1 AA review of UI artefacts: contrast, semantics, focus, ARIA. Trigger per design review pass.
ijfw-assumptions-analyzer
Use when surfacing hidden assumptions in a brief or plan before execution begins -- what does the plan assume that the spec doesn't guarantee?
ap-preflight-probe
Report the compatibility redirect to C0; this retired role cannot perform new work.
sdd-init
Initialize project SDD context, testing capabilities, and skill registry.
architect
System design, scalability analysis, and technology decision-making.