AGENT SKILL

project-optimize

暂无简介

License 许可
Apache-2.0
所属仓库
iokaio/vcp
最近更新
2026-10-06
Stars
★ 0
Forks
0

02 资源文件 1

03 SKILL.md 正文

Project routing and context optimization

Original VCP guidance.

Inspect first

  • VCP's retained evidence for the project: /optimize report and /optimize status in the terminal (offline: vcp optimize report --from <ms> --until <ms> and vcp optimize status), task cost records, latency observations and routing outcomes for the window under review.
  • Current routing configuration and policy: selected model/endpoint/group restrictions, project pins, output and input limits, escalation settings, and the developer-selected quality floor.
  • Trusted limits that cap any request: budgets, ceilings and allowlists outside project control.
  • Which report fields are known spend, reserved or unresolved liability, estimates, and independently graded quality versus verification observations.
  • The sample size and age of the evidence, and whether cohorts are comparable.

Inspect the evidence

Use the project's existing /optimize report/status workflow and current routing configuration. Compare like task cohorts and report windows. Distinguish known spend from unresolved liability, estimated costs from actual charges, selector timing from model latency, and verification observations from independently graded quality.

Ask only for preferences the report cannot establish, recording answers with /optimize answer priority|size|review|restrictions <value>. Priority, expected size, review needs, and model restrictions are input to an explicit proposal, not permission to widen trusted allowlists or spend evaluation money. Do not infer a numeric quality floor from prose or treat a synthetic benchmark as live qualification.

Proceed and verify

  1. Summarize the retained evidence by cohort and window, and name the gaps: too few samples, stale data, mixed task types, or unresolved costs.
  2. Prepare a concrete preview with /optimize preview low|med|high --quality-floor <bps>, using the developer-selected quality floor in basis points. Explain requested and effective policy under trusted limits.
  3. Propose, do not silently change. Budgets, quality floors, allowlists and routing policy change only when the developer selects a specific current preview.
  4. Apply only an explicitly selected current preview with /optimize apply; stale revisions require refresh. Rollback must identify target and expected current revisions: /optimize rollback <target> --expected <current>.
  5. After an applied change, compare later retained evidence against the same cohort definition before claiming an improvement, for example with /optimize compare <baseline-report> <current-report>.

Evidence is the report window, cohort definition, the preview identity that was applied, and subsequent observations. A preview is a proposal; an applied change without later observations is not a demonstrated saving.

Pitfalls

  • Comparing cohorts with different task mixes, sizes or review requirements.
  • Counting only successful attempts and hiding retries, escalations, review or support costs.
  • Treating an estimate or reservation as an actual charge, or ignoring uncertain cost.
  • Lowering a quality floor or skipping required review to make cost look better.
  • Starting paid model trials or remote advice to fill evidence gaps without explicit budget authority.
  • Reducing context in ways that drop required capabilities, instructions or source attribution.

Results

Report evidence gaps, proposed changes, expected trade-offs and subsequent observed results. Context reductions must preserve required capabilities and source attribution. Never hide retries/support costs or start paid model trials to make a comparison look complete.

Authority: this guidance ranks below current user constraints and AGENTS.md, grants no tools, installs or network access, and missing prerequisites are reported as not run.

04 同仓库更多 Skills

iokaio
iokaio/vcp

skill-installer

Install Codex skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo (including private repos).

★ 0
iokaio
iokaio/vcp

skill-creator

Create or update a Codex skill with appropriately scoped instructions and any needed supporting resources.

★ 0
iokaio
iokaio/vcp

review-agent

Perform a read-only, defect-first review of a specified code change and return every actionable finding. Use when another agent delegates review of uncommitted changes, a base-branch diff, a commit, or custom review instructions.

★ 0
iokaio
iokaio/vcp

plugin-creator

Create and scaffold plugin directories for Codex with a required `.codex-plugin/plugin.json`, optional plugin folders/files, valid manifest defaults, and personal-marketplace entries by default. Use when Codex needs to create a new personal plugin, add optional plugin structure, generate or update marketplace entries for plugin ordering and availability metadata, or update an existing local plugin during development with the CLI-driven cachebuster and reinstall flow.

★ 0
iokaio
iokaio/vcp

openai-docs

Use for Codex models/pricing, scheduled tasks, skills, settings, setup, troubleshooting, customization, automations, and self-knowledge—including 'you,' 'your,' 'this app,' or 'this coding agent' when they refer to Codex—and for OpenAI APIs/products and ChatGPT Work. Also use for model choice/migration, prompting, SDKs, Responses, Realtime, agents, evals, and Chat/Work/Codex comparisons. Do not use for generic app/software tasks that merely mention Codex.

★ 0
iokaio
iokaio/vcp

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG/vector/code-native assets, extending an established icon or logo system, or building the visual directly in HTML/CSS/canvas.

★ 0