This repository was archived by the owner on Jul 19, 2026. It is now read-only.
docs: self-improvement core — ADRs + thin index#12
Merged
Conversation
…le-responsibility agents) Design for airis-code's core: single-responsibility agent roster (operator/ test-gate/implementer/verifier/independent-reviewer/researcher + passive telemetry), enforced via per-role model pins + structured-output gates + PreToolUse deny hooks; an eval-driven evaluator-optimizer self-improvement outer loop gated on measured quota/quality/rework/speed. Cross-platform common-core + Claude/Codex adapters. Grounded in official Anthropic/OpenAI docs with sources; sequenced Phase 0-3 (telemetry persistence first). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Restructure per documentation best practices (ADRs in source control, MADR v4.0.0, immutable/supersede-only): decisions -> docs/adr/0001-0004; the former monolithic design doc -> a thin index linking ADRs + issues. Decisions now resist rot (ADRs are never edited, only superseded); work lives in issues (#13 = Phase 0); contracts will live in code + tests. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
kazukinakai
changed the title
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 subscribe to this conversation on GitHub.
Already have an account?
Sign in.
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Records the design for airis-code's self-improvement core as immutable ADRs (MADR v4.0.0, in
docs/adr/) plus a thin index — chosen over a standalone design doc because design docs rot (they must be "kept up to date"), whereas ADRs are immutable/supersede-only and issues are ephemeral.Contents
docs/adr/0001single-responsibility agent rosterdocs/adr/0002enforce role boundaries mechanically (per-role model pin + structured-output gates +PreToolUsedeny hooks)docs/adr/0003optimize for the subscription quota ceiling, not token thrift (operator stays frontier = quota-optimal)docs/adr/0004measurement-first, eval-driven self-improvement; common core + Claude/Codex adapters; strict Phase 0→3docs/adr/README.mddecision-log index + immutability conventiondocs/SELF_IMPROVEMENT_CORE.mdreduced to a thin overview/indexRationale (sourced)
Decisions → immutable ADRs (never edited, only superseded → can't go stale); work → GitHub Issues (Phase 0 = #13); contracts → code + tests (living). Cost/quota claims backed by the confidence-tagged evidence dossier. No code changes — decision records only.
🤖 Generated with Claude Code