brenbuilds1
- 7 skills
- 0 followers
- 6 hours ago last updated
- ▌ Unclebob · brenbuilds1 bundleA do-not-read-the-code workflow for delegating implementation to a coding agent. Use when handing an agent a feature or a fix you do not intend to review line by line. The agent turns the request into human-approved gherkin scenarios, writes failing tests and locks them before writing any implementation, then runs the result through machine-checked constraints: unit tests, acceptance scenarios, changed-line coverage, mutation testing on both the code and the gherkin examples, quality metrics, the project's own checks, plus a torture round (property, performance, jitter tests) where the change warrants it. The human reads a one-table receipt instead of the diff. The test lock and the mutation constraint exist because the same agent writes both the code and the tests.
- ▌ Ste Audit · brenbuilds1 bundleCheck output against the machine-checkable ASD-STE100 rules: sentence length caps, noun cluster limits, allowed verb forms, -ing restrictions, passive voice, paragraph limits. Use after adding an STE memory line, CLAUDE.md rule, or writer skill, or when replies drift wordy again. Violations table with Issue 9 rule numbers and quoted receipts, plus a drift verdict (holding, drifting, gone). Reports only; never rewrites; no dictionary bundled.
- ▌ Loop Referee · brenbuilds1 bundleReferee for writer/reviewer loops between coding agents, where one tool writes code and the other reviews it in rounds: Claude Code writing while Codex reviews, or the reverse. Use when running or setting up a cross-model review loop, a test-fix loop, or any "loop until clean" workflow. Sets the loop contract up front, keeps an append-only per-round ledger, applies stop rules so the loop converges, ends, or escalates to a human instead of burning tokens on two agents disagreeing. Does not write or fix code itself.
- ▌ Memory Audit · brenbuilds1 bundleFact-check an agent's persistent memory against the repo it describes. Use when an agent acts on outdated notes, when inheriting a project with existing CLAUDE.md / AGENTS.md / rules files, after a large refactor, or periodically on any long-lived project. Inventories every memory and instruction file the agent loads, splits them into atomic claims, verifies each claim against the current codebase and git history, and reports a per-line verdict table (confirmed, stale, wrong, contradiction, preference) with a receipt for every verdict. Reports only; never edits or deletes memory.
- ▌ Render Critic · brenbuilds1Use when agent-built 3D looks flat, sloppy, or AI-made: a three.js or WebGPU game scene, a Blender render, a glTF export, a Unity or Unreal level. Also use when a make-it-look-better loop keeps circling, when animation looks floaty or feet slide, when self-graded scores inflate while the scene stays ugly, or before shipping a scene people judge by eye.
- ▌ Skill Auditor · brenbuilds1 bundleAudit a third-party agent skill (Claude Code / Codex / Gemini CLI SKILL.md packages) before you install it. Use when asked to audit, vet, review, or security-check a skill, plugin, or marketplace package, or before installing one from GitHub or a marketplace. Clones the package read-only, runs supply-chain greps for fetch-and-execute, credential access, hidden code, network calls, and broad permissions, then reports findings with exact file:line and an honest severity for each. Does not install anything.
- ▌ Engineering Review · brenbuilds1 bundleCode review stance for serious engineering changes. Use when asked to review a diff, PR, commit, branch, patch, architecture change, migration, implementation, or test plan for bugs, regressions, missing tests, operational risk, security/privacy issues, and maintainability problems that could actually hurt users.