Test multi-sub-agent Review Agent - #3
Merged
ben-alkov merged 13 commits intoJun 2, 2026
Conversation
The log messages in submitFormalReview() said "finding(s) omitted" when out-of-diff findings could not be posted as inline comments. This wording led agents to incorrectly conclude the findings were irrelevant to the verdict. In reality, the GitHub API cannot accept review comments on lines outside the PR diff — the findings themselves remain valid and still influence the verdict. Changes: - postreview.go: Reword log messages from "finding(s) omitted: ..." to "inline comment(s) omitted (...) — findings still count toward verdict". Add a code comment explaining the GitHub API limitation. - postreview_test.go: Update test assertions to match new messages. - pr-review/SKILL.md: Add a "Findings vs inline comments" section documenting the pipeline: findings determine the verdict; inline comments are a delivery mechanism with GitHub API constraints. Note: pre-commit could not run in sandbox (Go toolchain permission error, exit 3). TestResolveLinuxBinary_Download failed due to sandbox network restrictions (pre-existing, unrelated). Closes fullsend-ai#1537
ben-alkov
force-pushed
the
feat-review-agent-parallel-specialized-subagents
branch
3 times, most recently
from
May 28, 2026 20:00
5429f13 to
91af2bd
Compare
The botsend account used by e2e tests accumulates classic PATs over time when test runs crash or time out before cleanup. Add a Playwright- based script that paginates the GitHub tokens settings page, deletes all expired tokens, and reports how many remain. - hack/cleanup-pats: bash wrapper for manual use - hack/cleanup-pats.go: Playwright Go program that deletes expired PATs via fetch+POST against each token's delete form - .github/workflows/pat-cleanup.yml: weekly cron (Sundays 4am UTC) using the existing E2E_GITHUB_SESSION secret Closes fullsend-ai#1683 Assisted-by: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Ralph Bean <rbean@redhat.com>
ben-alkov
force-pushed
the
feat-review-agent-parallel-specialized-subagents
branch
2 times, most recently
from
May 29, 2026 20:01
de617ba to
4224b6d
Compare
Introduces internal/resolve package that orchestrates fetch, cache, validation, and audit logging for URL-referenced declarative harness fields (agent, policy, skills). Modifies the harness in place, replacing URLs with local cache paths. Phase 1: single-level only. Moves NewTestPolicy to internal/fetch/fetchtest/ package for safe cross-package test use, and fixes ResolveRelativeTo to skip URL-valued fields. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Greg Allen <gallen@redhat.com>
Remove hack/cleanup-pats shell script and call go run hack/cleanup-pats.go from the workflow step. Assisted-by: Claude claude-opus-4-6 <noreply@anthropic.com> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Ralph Bean <rbean@redhat.com>
The previous commit deleted the bash wrapper but missed updating the workflow step that invoked it. Assisted-by: Claude claude-opus-4-6 <noreply@anthropic.com> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Ralph Bean <rbean@redhat.com>
ben-alkov
force-pushed
the
feat-review-agent-parallel-specialized-subagents
branch
from
June 1, 2026 14:59
4224b6d to
cfbe17f
Compare
Single-pass monolithic review cannot scale depth with PR complexity. Specialized sub-agents let the orchestrator fan out independent dimensions concurrently, each with model pinning tuned to its task. Assisted-by: Claude Code (Opus 4.6) Signed-off-by: Ben Alkov <ben.alkov@redhat.com>
Single-pass review misses domain-specific issues and cannot scale review depth with PR complexity. Orchestrator pattern enables parallel specialist dispatch across nine review dimensions. Assisted-by: Claude Code (Opus 4.6) Signed-off-by: Ben Alkov <ben.alkov@redhat.com>
The orchestrator dispatches docs-review as a sub-agent, but docs-review also dispatches its own sub-agent. Without context detection via REVIEW_SUB_AGENT_TRUE, this creates wasteful nested dispatch. Assisted-by: Claude Code (Opus 4.6) Signed-off-by: Ben Alkov <ben.alkov@redhat.com>
ben-alkov
force-pushed
the
feat-review-agent-parallel-specialized-subagents
branch
from
June 1, 2026 15:07
cfbe17f to
0681c7d
Compare
Assisted-by: Claude Code (Opus 4.6) Signed-off-by: Ben Alkov <ben.alkov@redhat.com>
ben-alkov
force-pushed
the
feat-review-agent-parallel-specialized-subagents
branch
from
June 1, 2026 15:24
0681c7d to
0c8944b
Compare
…solver feat: add resource resolver for URL-referenced harness resources
…parallel-specialized-subagents feat(review): parallel specialized sub-agents for PR review
…fy-findings-vs-comments docs(fullsend-ai#1537): clarify findings vs inline comments distinction
…nd-pats ci: add weekly cleanup of expired e2e PATs
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
3 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.
No description provided.