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.
npx agentmods add skills/nickzaitsev/ru-normalizr/implement-plannpx skills add NickZaitsev/ru-normalizr --skill implement-plangit clone --depth 1 https://github.com/NickZaitsev/ru-normalizrWrote 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.
[](https://agentmods.dev/skills/nickzaitsev/ru-normalizr/implement-plan)<a href="https://agentmods.dev/skills/nickzaitsev/ru-normalizr/implement-plan"><img src="https://agentmods.dev/badge/skills/nickzaitsev/ru-normalizr/implement-plan.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00048 | $0.00723 |
| Opus 5 | $0.00024 | $0.00362 |
| Sonnet 5 | $0.00010 | $0.00145 |
| Haiku 4.5 | $0.00005 | $0.00072 |
Grade A, and why
implement-plan 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 6d 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.
How it starts
The opening of the file, as written. The whole thing — 35 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Languages: all edits to the plan file are in English (like the plan itself). ALL communication with the user — questions, discrepancy reports, the final report — is in Russian.
Process
- Locate and read the whole plan. If the user didn't name a file, list plans in
docs/agents/plans/with statusplan/in progressand ask in Russian which one to implement. If the plan's status isimplementedorrejected, stop and confirm with the user before proceeding. - Mark the plan started: set
Status: in progress since YYYY-MM-DD.. - Implement step by step in the plan's order, committing per phase (atomic conventional commits per repo rules).
- Test as you go, but don't over-test. Run the plan's checks per stage, BUT skip a stage's test run when a later plan stage will exercise the same code anyway — speed matters; one honest end-to-end verification at the end (plan's "Testing & verification" section + repo lint/type-check/test rules) is mandatory and cannot be skipped.
- After each large stage (or a few small ones), compact the context yourself if your harness lets you trigger compaction. Stage boundaries are where compacting loses the least.
- Close the plan — this is part of "Done":
- Update the status line:
Status: implemented YYYY-MM-DD in branch \`.` (add short caveats if something was skipped). - Correct any sections where the implementation diverged — the plan must not lie about what was built.
- Commit the plan update (may be folded into the final feature commit).
- Update the status line:
- Report to the user in Russian — a TL;DR, not a wall of text: a few short bullets covering what was built, any divergences from the plan, and how it was verified. Details stay in the plan file and commits; expand only if the user asks.
If some change doesn't make sense or wrong, let the user know.
Don't execute a broken step
No dedicated verification pass is needed — plans are usually implemented right after being written. But keep your eyes open: if while implementing you see that a plan step is wrong or won't work (relies on code that doesn't exist / behaves differently, contradicts something you've already built, would produce a broken result), do NOT execute it as written. Stop, explain in Russian what's wrong and what you propose instead, and agree with the user before continuing.
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.
- 6d ago First seen · 35 lines · 48 tokens per session scan A 3c4f4b37dcff
implement-plan is a skill published in the GitHub repository NickZaitsev/ru-normalizr (22 stars, last pushed 29d ago), licensed MIT. It adds 48 tokens to every session and 723 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
api-contrib
Contributing to the Kokoro-FastAPI Python API: module layout, endpoint gating pattern, test expectations. Use when adding or changing endpoints, services, or inference code.
add-migration
Write a migration ledger entry for a configuration schema change in pipelex, so that a user's existing file can be repaired instead of thrown away. Use when make check-migration-schemas (alias cmig) or make check-ledger (alias cl) refuses a change, when the user says "add a migration", "write the ledger entry", "the…
bump-kajson
Move this repo's exact kajson dependency pin to the latest release on PyPI (or a version you name), re-lock, adapt the engine to whatever the new release changed, retire any workaround the release makes unnecessary, run the checks, and write the CHANGELOG entry — stopping before the commit. Use whenever the user says…
update-graph-ui
Bump the pinned @pipelex/mthds-ui (and elkjs) version that the generated ReactFlow HTML loads from jsDelivr. Re-fetches the bundle, recomputes SRI hashes, and updates the Python constants in standaloneassets.py. Use when user says "update graph ui", "bump mthds-ui", "update graph viewer", "new version of mthds-ui", or…
minimax-cli
Nested swiss-knife reference for the MiniMax mmx CLI and the canonical MiniMax CLI procedure shipped with the TUI: install mmx-cli, discover the correct TUI-managed MiniMax preset/key slot without leaking secrets, match mainland vs international regions, and route image/video/music/TTS generation or one-shot shell…
video-script
对已完成分析的视频进行导演与剪辑策划，再写带时间戳的中文解说并校验。workdir 已包含 agentnarrationbrief.md 与 vlmanalysis.json 时使用。适用于故事方向、片段选择、画面/原声/旁白分工、 解说写作与复核。输入 workdir 中的理解索引；输出 recapstoryplan.json、visualaudioboard.json、 可选 stylecard.json、cut 模式需要的 clipplan.json，以及通过校验的 narration.json。触发词：解说词、写解说、视频旁白、 narration script、写稿、解说文案、剪辑思路、导演思路。.