agentleFS
Fadi88Fadi88/one-man-startup/.claude/skills/cost-audit/SKILL.md

cost-audit

Skill16 starsChanged 17 days agoMITView on GitHub

The optimizer's method for finding cost and waste in the agent workflow - model tier per agent, context loaded per step, redundant and skipped steps, slow or useless checks - and proposing ranked, reversible cuts. Use monthly, on a spend jump, or when the founder asks why something was slow or expensive.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Checks: run time per check from the sweep output.

    line 19
  2. Very specific:

    Never propose removing a gate, QA, security or legal review.

    line 41

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<!-- Generated by tools/sync_adapters.py from skills/cost-audit/SKILL.md. Edit the source, not this file. -->

Cost audit

Gather (read only)

  1. Spend: whatever usage export the founder's tools give (per day, per model). If none exists, say so and estimate from docs/work/*/status.md (agents per feature, rounds per step), labelled as an estimate.
  2. Shape: for the last few features, which agents ran, how many rounds each step took, how many times work came back from qa.
  3. Context: the size of AGENTS.md, every agent and skill file, and the documents agents read most. python tools/run_all.py prints the files over budget.
  4. Checks: run time per check from the sweep output; checks that have not failed in months.

Look for

  • Savings counted before they were measured, such as prompt caching assumed in a cost table. Caches are specific to one provider and expire, so failover and low traffic both erase them.
  • An agent on the deep tier doing checklist work, or a fast-tier agent whose work keeps coming back from review (too cheap is also waste).
  • Work that went down the feature track and should have been a fix, or the reverse.
  • Reviews repeated with no new input; parallel agents whose output was not used.
  • Artifacts passed by content instead of path; whole files read where a grep would do.
  • Shared files grown past budget.
  • Rework: the costliest waste is a build that fails QA because the spec was thin. Sometimes the saving is more grilling, not less.

Propose

docs/cost/<YYYY-MM-DD>.md, one table, ranked by saving:

ChangeSaving (measured or estimated)Risk to qualityHow to undoHow we will know in two weeks

Never propose removing a gate, QA, security or legal review. Proposing a cheaper way to run one is fine. Each accepted proposal becomes an experiment (skills/experiment/).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.