feat(tools): #1246 #1247 #1248 — coding-agent feedback surface (research, tool guidance, ruff diagnostics)#1261
Open
ohdearquant wants to merge 2 commits into
Open
feat(tools): #1246 #1247 #1248 — coding-agent feedback surface (research, tool guidance, ruff diagnostics)#1261ohdearquant wants to merge 2 commits into
ohdearquant wants to merge 2 commits into
Conversation
…nostics Give the coding agent an IDE-grade feedback loop across reader/editor/bash and a new static-analysis check tool. - #1246: research OSS harnesses (OpenCode, mini-swe-agent) — agent loop, tool set, permission model, persistence. docs/research/coding-harnesses-2026-06.md with pinned commit SHAs, file:line citations, and a prioritized fold-in list. Drafts the remaining AST surface for the #1247 follow-up. - #1248: tune reader/editor/bash guidance — error messages now explain why a failure happened and how to recover (failed edit → line-prefix/whitespace diagnostics + re-read; bash → cwd= over `cd &&`, PATH, timeout, truncation). Tests assert the recovery text and that recovery paths work. - #1247: AST/static-analysis feedback tool — code_check shells out to ruff (optional dep, shutil.which guard, never raises) returning structured file:line:col diagnostics. Composes with the editor (edit → check). Test exercises a known-bad snippet end to end. Tests: uv run pytest tests/tools/test_check.py tests/tools/test_guidance.py tests/tools/test_reader.py — 63 passed. Critic verdict: APPROVE (CRIT:0 MAJ:0 MIN:3). Refs #1246 #1247 #1248 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sweep results: research tool, tool guidance, ruff diagnostics modules plus codebase-wide UP038/formatting fixes applied by the sweep. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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
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.
Coding-harness feedback-surface slice from the
lionagi-sweepshow — the SWE-bench engagement thesis (harness > model). Critic APPROVE (CRIT:0 MAJ:0 MIN:3), claims verified against source; 63 tests pass; pre-commit ruff green.Shipped
docs/research/coding-harnesses-2026-06.md: concrete findings from OpenCode + mini-swe-agent (loop, tool set, permission model, persistence) + a prioritized "fold into lionagi" list.tools/file/reader.py,tools/file/editor.py,tools/code/bash.py)code_check: a ruff-as-feedback diagnostic tool returning structuredfile:line:diagnostican agent can act on, composable with the editor. (tools/code/check.py)Deferred (drafted in the research doc)
:118-162) but not yet implemented. Leaving feat: AST/static-analysis coding tools — give agents IDE-grade feedback #1247 open for that follow-up.Closes #1246
Closes #1248
Refs #1247 (first slice shipped; AST remainder drafted)
🤖 Generated with Claude Code