← all publishers

Nowely

@nowely source repo

Report this profile Claim this profile

• 10 skills
• 0 followers
• 1 hour ago last updated

1. ▌
   Experiment · nowely bundle

   Runs one registered experiment on the orchestrator's own rules: arms under the orchestrate mode on matched material, a judge that does not see which arm it reads, the orchestrator's conclusion and the user's verdict, recorded under the plugin's data directory by a script and copied into the repository later.
   0 installs
2. ▌
   Orchestrate · nowely bundle

   User-invoked mode that turns this conversation into an orchestrator: scout inline, agree one plan, then push every verbose step (tests, tree-wide greps, source files, diffs, logs) onto Claude and Codex agents so the main context stays small.
   0 installs
3. ▌
   Calibrate · nowely bundle

   Measures which forms of writing one person prefers, by blind paired choice on a balanced item bank, and turns the result into rules a writer can follow and a judge score that is honest about its own error. Run across several sessions; re-run when it stops predicting.
   0 installs
4. ▌
   Audit · nowely bundle

   Measures a document against two rulers: whether fresh readers get the right answer, and whether every claim about behaviour is true of the code. Returns a reader profile, a claim ledger, reader scores and the list of what broke. It never proposes wording; `rewrite` does that.
   0 installs
5. ▌
   Codex · nowely bundle

   Delegates tasks to Codex as a subagent with per-call rights: analysis that writes nothing of yours, or writing and tests in a managed git worktree, each reaching the network unless the call denies it. Use when a panel, refuters, or competing designs need an agent that does not share Claude's bias; when fanning out reviewers or adversarial verifiers; after two hypotheses fail; when a second independent implementation is wanted; or when the user names Codex, GPT, or "the other model" (через codex, через gpt, вторая имплементация, панель ревьюеров), or names a Codex model by its short name (Astra, Sol, Terra, Luna; астра, сол, терра, луна). It also governs requested mixes ("one of them codex", "half codex", "only codex") and refusals ("no codex", "just you"). Skip trivia and mechanical fact-gathering.
   0 installs
6. ▌
   Swarm · nowely bundle

   Runs a swarm: up to fifty bulk agents over a file of units, launched by one script with a concurrency cap, and reduced by one cheap agent to a single summary the orchestrator reads; shared state and free messaging between agents are experiment arms, never the default.
   0 installs
7. ▌
   Rethink · nowely bundle

   Decides what a document should be before a sentence of it is written: what comparable documents already solved, what things are called, and what is said in what order. Returns a skeleton and stops there — the writing is `rewrite`'s. Use when a document's shape is wrong, or when starting one.
   0 installs
8. ▌
   Rewrite · nowely bundle

   Writes the text in rounds: one candidate, then critics with lenses that differ, then edits declared with the check behind each, until the owner reads a round and says whether they would send it as it is. Starts from a skeleton `rethink` agreed, or from the failures an `audit` measured. Proposes; writes into your tree only on your word.
   0 installs
9. ▌
   Advisor · nowely

   Adds a standing advisor to one orchestrated run: a top-row agent of the other model family, kept for the run and asked one question at each decision point; the run records every decision point and what the advice changed, because the advisor is an experiment until measured.
   0 installs
10. ▌
    Cleanup · nowely

    Lists files left by entrust, suggests what to remove, and deletes the user's selection after approval.
    0 installs