feat(harness): registry-grounded system prompt with coder routing and SDK-doc gate#245
Merged
Merged
Conversation
… SDK-doc gate Extend the turn-orchestrator identity prompt (all four provider variants) from engine-only discovery to the full capability ladder: registered function -> install from the public registry -> author a worker. - Registry flow: search directory::registry::workers::list/info, announce, install via worker::add, re-verify via prefix list; bootstrap or degrade when the directory worker is absent. Both catalogue calls are documented in-prompt and exempt from the fetch-contract-first rule. - Worked example grounded in the live registry: the published "email" worker (email::send) replaces the unpublished "resend"; the example now models the worker::add contract fetch and a plain announce line. - Coder routing: create/edit/move/delete code files go through coder::*, including coder::move for renames (never delete-then-recreate); the enumeration is non-exhaustive with the prefix list as inventory, and the shell::fs boundary is drawn for non-code browsing. - SDK-doc gate: fetch the per-language iii.dev SDK reference before the first line of worker code, scoped to new workers/registrations, with llms.txt recovery, graceful degradation, and the ordinary-call guard. - web::fetch mandate extended to localhost/just-bound triggers with an explicit success criterion (read ok/status/body), no curl exception. - Tests: 146 pins incl. a directory::* allowlist invariant (registry catalogue only), coder::move, capability-ladder ordering, and the email example (mutation-tested as load-bearing). - Permissions: allow the read-only ids the prompt mandates (registry catalogue reads, read-only coder surface, web::fetch); worker::add and mutating coder ops stay approval-gated. Architecture/worker docs updated to the engine-grounded framing.
Contributor
skill-check — worker0 verified, 15 skipped (no docs/).
Four for four. Nicely done. |
|
Caution Review failedPull request was closed or merged during review No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (9)
📝 WalkthroughWalkthroughThis PR shifts the iii agent turn-orchestrator from a self-contained to an "engine-grounded" system: agents discover capabilities from the live engine and public registry, route code operations through a specialized ChangesEngine-grounded agent system with registry discovery and specialized worker routing
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
sergiofilhowz
approved these changes
Jun 11, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Extends the turn-orchestrator identity prompt (all four provider variants: anthropic, gpt, kimi, default) from engine-only discovery to the full capability ladder: registered function → install from the public registry → author a worker. Hardened by a 5-dimension adversarially-verified review of the draft; all confirmed findings are fixed in this PR.
Prompt changes (byte-equivalent semantics across all 4 variants)
directory::registry::workers::list/info→ announce →worker::add { source: { kind: "registry", name } }→ re-verify viaengine::functions::list { prefix }. Bootstraps a missing directory worker (iii-directory) and degrades gracefully when the registry is unreachable. Both catalogue calls are documented in-prompt and exempt from the fetch-contract-first rule (resolving a contradiction with RULE 2).emailworker (email::send); the previous draft modeledresend, whichGET /w?search=resendreturns empty for today. The example now also models fetching theworker::addcontract and announcing the install as a plain assistant line.coder::*, includingcoder::movefor renames (never delete-then-recreate). The function list reads as non-exhaustive with the prefix list as the inventory, there is a post-install re-check, and theshell::fsboundary is drawn for non-code browsing.llms.txtrecovery on fetch failure, graceful degradation when docs are unreachable, and the never-fetch-docs-for-an-ordinary-call guard.web::fetch(no curl exception), with an explicit success criterion: read the envelope (ok,status, body matches the handler).Tests (142 → 146)
directory::*allowlist invariant replaces the substring bans: everydirectory::id in the prompt must start withdirectory::registry::workers::(also blocks the legacydirectory::engine::*doc proxies).coder::move, non-exhaustive inventory + no delete-then-recreate, ordinary-call guard, degradation,.mdtrick, engine-sdk fallback, llms.txt recovery, capability-ladder ordering (registry → coder → authoring), and the email example (mutation-tested: deleting the block fails the suite).Permissions + docs
iii-permissions.yaml: allow the read-only ids the prompt now mandates — registry catalogue reads, read-onlycoder::*surface, andweb::fetch(size/timeout caps + server-side SSRF protection; load-bearing for the SDK gate and trigger verification).worker::addand mutatingcoder::*ops stay approval-gated.harness/docs/architecture.mdallow-rule enumeration updated to match; stale "engine-only / discovers everything from the live engine" framing replaced with "engine-grounded" inturn-orchestrator.mdand thesystem-prompt.tsJSDoc.Test plan
pnpm vitest run tests— 1371/1371 pass (122 files), including the policy suite against the edited yamlnpx tsc --noEmitclean;biome checkclean on touched files<example>block byte-identical across all 4 variantsemailworker verified published withemail::send;resendconfirmed absentweb::fetch(not curl) for trigger verification and fetches the SDK reference before authoringSummary by CodeRabbit
Release Notes
Documentation
New Features
Tests
Chores