Skip to content

kklimuk/docx-cli / .claude/skills/weak-agent-test

weak-agent-test

Verified

Run the weak-agent adversarial test harness against docx-cli. Spawns weak exercise agents (Haiku by default, Sonnet to probe, or a local agent harness's pre-produced runs) to perform real document tasks over six scenarios — five editing (MNDA form-fill + font fidelity, invoice table-edit/restructure + logo replace, résumé styling, contract redlining + commenting, contract finalize via accept/reject + comment reply/resolve) and one authoring (T. S. Eliot poetry journal: multi-column, verse, footnotes, links, figure) — renders every result with Word, has opus judge them against ground-truth rubrics, measures each exercise's tool economy, token cost, wall-clock, and correctness (from transcripts for Claude, the exercise.json ledger for the local harness), and synthesizes a prioritized ergonomics report. Use when the user says 'adversarial review', 'test docx-cli with weak agents', 'run the haiku harness', 'weak agent test', or wants to re-run yesterday's adversarial process.

scripts/local-exercise-prompt.md

View on GitHub
You are testing **docx-cli**, a command-line tool for reading and editing Microsoft Word (.docx) files. You have NOT used it before — discover what you need from its own `--help`; do not assume flags.

You have a **bash** tool. To read a text file, run `cat <path>`. To run docx-cli, invoke the executable at this exact path:
  {{BINARY}}

First, orient yourself by running these with bash:
  {{BINARY}} --help
  {{BINARY}} info locators
  {{BINARY}} <command> --help     (for any command you decide to use)

Also list your assets folder (it may be empty); if it holds files, read them:
  ls assets

{{WORKLINE}}
  {{DOC}}

## Your task

{{TASK}}

## Your job
Carry out the task above on the working document. The request describes the OUTCOME the person wants — YOU work out which docx-cli commands get there.

## Rules
- Use ONLY the docx-cli executable above for document operations. Do NOT unzip the .docx, hand-edit XML, or use any other tool for the document.
- STAY IN YOUR CURRENT FOLDER. Do not read or list files outside it — no `ls -R /`, no exploring parent directories. Everything you need is here.
- Locators like p0, t0:r1c2:p0, sN shift after structural edits — re-read the document (`{{BINARY}} read {{DOC}}`) when unsure. Prefer batch commands where the tool offers them.
- If a command fails, read its error, try a couple of sensible fixes, then move on. Do not loop on one step forever.
- Finish the task if you can.

When you are done, write ONE short paragraph summarizing what you changed.
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 321kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 321kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 321kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 197kUpdated Claude CodeCodexVerifiedInstall