Fan out subagents for socratic-review's silent assessment - #32
Merged
Conversation
Step 0 previously read all code inline on the main thread. For large or unfamiliar targets (multi-file PRs, SHAs, inherited code), dispatch one subagent per problem space in parallel, then merge their findings into the private ranked list. Small targets are still read inline to avoid overhead. The assessment stays silent — subagent reports are never surfaced. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This was referenced Jun 12, 2026
Merged
stevepolitodesign
added a commit
that referenced
this pull request
Jul 29, 2026
Step 0 has fanned out to subagents since #32, but the instruction was vague enough to be unreliable. It said only "dispatch subagents" — naming neither the tool nor the agent type — so how the assessment actually got parallelized varied from session to session. The bigger gap was what happened when the subagents came back. Nothing told the reviewer to read the code itself, and it needs to: Step 1 opens by naming what it sees, Step 2 leads the user into specific blind spots, and Step 3 guides real refactoring moves. None of that is possible from a list of smells with severity labels and line numbers. The per-subagent brief now lives in `references/smell-explorer.md`, following the pattern feature-dev established. It asks for the evidence behind each smell, not just its name, and it forbids review prose so the orchestrator keeps control of what surfaces and when. Co-Authored-By: Claude Opus 5 (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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Updates
socratic-review's Step 0: Silent Assessment so that large or unfamiliar review targets are explored by parallel subagents instead of read entirely inline on the main thread.Why
An audit of the skills found
socratic-reviewwas the one skill doing substantial codebase analysis inline with no subagent isolation.explainandprior-artalready run as forkedExploreagents (context: fork), butsocratic-reviewcan't use that pattern — it's an interactive Socratic dialogue and must stay on the main thread to ask one question at a time. So the fan-out lives inside Step 0 and returns control to the conversation.How
🤖 Generated with Claude Code