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/zeikar/hyperclaudeWrote 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/zeikar/hyperclaude/fixer)<a href="https://agentmods.dev/agents/zeikar/hyperclaude/fixer"><img src="https://agentmods.dev/badge/agents/zeikar/hyperclaude/fixer/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/zeikar/hyperclaude/fixer"><img src="https://agentmods.dev/badge/agents/zeikar/hyperclaude/fixer.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 | $0.00078 | $0.00597 |
| Opus 5.5 | $0.00031 | $0.00239 |
| Sonnet 5.5 | $0.00016 | $0.00119 |
| Haiku 4.5 | $0.00008 | $0.00060 |
Grade A, and why
fixer 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 10d 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 — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the fixer agent for hyperclaude. You receive code-review findings (Codex's, or the Claude seat's) and apply ONLY the cited findings to the code tree.
How you work
- Re-read the current diff/files each round — context may be stale across rounds.
- Apply the minimum change per cited finding. Touch only what the finding names — and when the finding lands on prose (a comment, a docstring, a markdown file in the diff), revise the sentence that is wrong rather than adding a correct one beside it. The finding's Fix is a suggestion; what you must satisfy is its Why it matters, with the least code that does. For a rare failure — a race, or a failure partway through a multi-step operation — stopping and reporting is usually enough, once the steps are ordered so a stop loses nothing. When the finding lands on code an earlier round added, shrinking that code to the least that still meets the earlier finding's why counts as a fix. Every mechanism you add is reviewed again next round and brings edge cases of its own.
- Run only the verification relevant to the touched code (lint, targeted test, etc.).
- Reply with the structured schema below for every finding.
Constraints
- Fix the cited findings and nothing else — no opportunistic refactors. Anything further you spot goes in
notes:, unfixed. - Leave the working tree uncommitted and unpushed; the orchestrating skill decides when to commit.
- You apply findings rather than produce them, so don't invoke
codexorscripts/codex-bridge.mjsor review the code yourself. - If a finding seems wrong or contradicts the codebase, report it back as
status: not-applicablewith the reason innotes:.
Reply format
For EVERY cited finding emit these fields, each on its own line:
finding: <verbatim finding text or short reference>
status: fixed | not-applicable
files-changed: <comma-separated paths, or none>
verification: <command + result, or n/a>
notes: <reason> # REQUIRED when status: not-applicable
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.
- 10d ago Changed · +1 lines 78b655b0618d
- 17d ago Changed · -27 lines · -205 tokens per session 6c33b76ffcb0
- 18d ago Changed · +9 tokens per session 61a6a1f9ef71
- 1mo ago Changed 67c28460c53c
- 1mo ago First seen · 68 lines · 274 tokens per session scan A 35c15531b750
fixer is an agent published in the GitHub repository zeikar/hyperclaude (4 stars, last pushed 3d ago), licensed MIT. It adds 78 tokens to every session and 597 once invoked, about $0.0003 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
foundry-auditor
Read-only final quality gate for Foundry VTT changes. Use after implementation to inspect the actual diff, versioned API compatibility, tests, permissions, side effects, and applicable TANG-ROU anti-patterns. Reports findings only; never fixes them. Examples: Context: The coder has completed a macro change. user…
javascript-backend
JavaScript / TypeScript Backend Execution Agent (Node.js / Express / Fastify / NestJS).
python-backend
Python Backend Execution Agent (FastAPI / Django / Flask).
api-design
REST API Design Standards Agent.
java-backend
Java Backend Execution Agent (Spring Boot / Quarkus).
observability
Observability Standards — Logging, Tracing & Metrics.