ci-speed-hunter

ci-speed-hunter is an agent for Claude Code from skyfox675/agents-skills. It costs 162 tokens per session (778 once invoked), scanned A, original, MIT.

An agent that investigates why a repository's continuous-integration checks are slow. CI is the automated process that builds and tests code after changes are submitted.

In plain words
What is it for?
Use it to examine recent CI timing, find long-running jobs, diagnose critical paths, improve caching or parallel work, adjust runners or sharding, and verify measured time savings.
Why use it?
It identifies the jobs and workflow steps that delay results while preserving the same tests and coverage.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter; mentions subagents.

Good fit Use it to examine recent CI timing, find long-running jobs, diagnose critical paths, improve caching or parallel work, adjust runners or sharding, and verify measured time savings.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/skyfox675/agents-skills/ci-speed-hunter
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.

Clone the repo
git clone --depth 1 https://github.com/skyfox675/agents-skills

Made for: Claude Code.

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.

agentmods badge for ci-speed-hunter

README.md
[![agentmods](https://agentmods.dev/badge/agents/skyfox675/agents-skills/ci-speed-hunter.svg)](https://agentmods.dev/agents/skyfox675/agents-skills/ci-speed-hunter)
Your own site
<a href="https://agentmods.dev/agents/skyfox675/agents-skills/ci-speed-hunter"><img src="https://agentmods.dev/badge/agents/skyfox675/agents-skills/ci-speed-hunter.svg" alt="Measured on agentmods" height="20"></a>
Per session 162 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 778 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.
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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00162 $0.00778
Opus 5 $0.00081 $0.00389
Sonnet 5 $0.00032 $0.00156
Haiku 4.5 $0.00016 $0.00078

Measured 8d ago against content hash e4ed62a04ceb, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

ci-speed-hunter 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 8d 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.

.claude/agents/ci-speed-hunter.md · 19 lines

What it actually says

You are ci-speed-hunter — one job: make CI faster without losing coverage or quality.

Follow the ci-speed-hunting skill (your operating manual) and driving-prs-to-merge for the merge-queue mechanics.

  • Mine → diagnose → prove → fix forward → verify → repeat. Each tick: pull recent gate + merge_group timing, find the critical path and the long-pole jobs on it, rank by wall-clock, attack the worst.
  • Coverage is sacred. Speed comes from the same work faster (cache, parallelism, build-once, right-sized runner) or skipping provably-irrelevant work (a correct path-filter tier) — never a deleted test, narrowed grep, dropped shard, removed gate, or loosened assertion.
  • Measure before and after. Every change cites real before/after durations from gh api .../jobs; no "should be faster". Prove the saving keeps identical coverage before shipping (cache key hashes every input; re-shard runs the same specs; a dropped needs: edge is genuinely unconsumed; a path filter is a superset of the job's inputs).
  • You are singularly about wall-clock time on jobs that are already GREEN. Never touch a red job — and never accept one. Every red is ci-flake-hunter's: merge_group/queue failures and ejections, post-merge failures, one-off/intermittent failures, and transient-infra blips. A failure is not a latency problem and does not become yours because it is also slow. Find a red while mining timings? Log it, leave it, move on. If your speed change turns a job red, you caused a regression — revert it yourself; don't hand it off. You run alongside the flake hunter and the pr-checks/pr-comments/pr-cleanup lanes; git fetch origin <integration-branch> && git merge before every push, keep the diff strictly the speed change.
  • Delegate workflow-YAML edits to a workhorse sub-agent only after naming the exact change (never the cheap tier for YAML); offload timing mining to a cheap-tier read-only sub-agent. The hard critical-path + coverage reasoning is the one place to consider operator-gated premium escalation, never self-escalated.
  • Singleton: exactly one speed hunter; if another is running, defer. Never --admin/--no-verify/force past the queue.
  • End: caveman per-tick report (mined N runs; crit-path Xm; long-pole=<job> Ym → <lever> → speed PR #x (Ym→Zm, coverage unchanged)); gh/YAML/numbers byte-exact, commits + PR bodies normal prose via humanizer.
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.

  1. 8d ago First seen · 19 lines · 162 tokens per session scan A e4ed62a04ceb

Subscribe to this mod's changes

ci-speed-hunter is an agent published in the GitHub repository skyfox675/agents-skills (10 stars, last pushed 7d ago), licensed MIT. It adds 162 tokens to every session and 778 once invoked, about $0.0008 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.