Repository navigation
chore: repo-compare.json: gate record against a MISS+failed-compare run producing no verdict - #637
Merged
Conversation
Owner
Author
|
Tracking this PR for automated review ( |
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (2)
Included review availability: 0 reviews are currently available. Your included PR review attempts over the past 7 days set your current allowance at 2 reviews per hour. 📝 WalkthroughWalkthroughThe workflow now evaluates recursive ChangesWorkflow condition updates
Estimated code review effort: 2 (Simple) | ~10 minutes Merge Risk: ⚪ Minimal · up to This change gates recording when a fresh comparison produces no verdict while preserving cache-hit behavior; no actionable merge-blocking risk remains after normal checks and review. 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Full details: Description checkExplanation The description provides the change summary, implementation details, test results, risk assessment, and compatibility context. It does not use the template headings or test-plan checkboxes, but it contains the required information and is mostly complete. Full details: Docstring CoverageExplanation Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 6 functions across 1 files. (1 skipped: 1 unsupported.) ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
…one.
**What I did:** Extended `compile_run_if` in `crates/flare-workflow/src/json.rs` to support a single top-level `OR`/`AND` combining two atomic conditions (each still `==`/`!=`/bare-truthy), and fixed a landmine I hit along the way: `expand_variables` leaves an unresolved `{{var}}` as literal template text rather than empty string, which would have made `verdict != ''` spuriously true whenever `compare` never ran. I now normalize any still-bracketed `{{...}}` output to falsy/empty inside `run_if` evaluation only (prompt expansion elsewhere is untouched, preserving the "agent sees the raw placeholder and refuses to fabricate" fallback).
Wired `record`'s gate in `.agentflare/workflows/repo-compare.json` to `{{cache_check}} != 'MISS' OR {{verdict}} != ''` — skips only on a fresh-analysis run with no verdict, still runs on a legitimate cache-hit passthrough.
**Tests:** added two unit tests directly against `compile_run_if` (`run_if_or_treats_unset_var_as_falsy_not_its_own_placeholder_text`, covering MISS+no-verdict skip / MISS+verdict run / cache-hit run; and `run_if_and_requires_both_sides`). Full `flare-workflow` crate suite: 29/29 passing, no regressions. JSON file validated as well-formed. Confirmed no other workflow file in the repo uses `run_if`, so this grammar change is isolated to `repo-compare.json`.
**Concerns:** the `OR`/`AND` split is a plain substring match, so a quoted literal containing the literal substring `" OR "` or `" AND "` would misparse — acceptable per the task's own scope note ("string-only, no full expression language"), but worth remembering if this grammar gets reused elsewhere.
Agentflare-Branch: task/181-repo-compare-json-gate-record-against-a
Agentflare-Item: 181-repo-compare-json-gate-record-against-a
getappz
force-pushed
the
task/181-repo-compare-json-gate-record-against-a
branch
from
August 31, 2026 06:43
f644eb4 to
9367452
Compare
1 task done
getappz
added a commit
that referenced
this pull request
Sep 24, 2026
…llow-up) CodeRabbit on #803 asked for a test exercising the dispatcher route rather than the compiled git shim directly. agentflare-shim's PATH dispatcher sets LEAN_CTX_AGENT before routing a command through lean-ctx -c git, but that marker was in neither agent-detector 0.2.1's catalog nor the AGENTFLARE_AGENT check, so a caller reaching this shim only via that route would misclassify as human and bypass the canonical-mutate guard entirely. - agent_invocation_detected() now also checks LEAN_CTX_AGENT directly. - human_shim strips it too, since it is ambient under lean-ctx sessions. - New test proves both branch-create and HEAD-detach stay denied under this marker, with the escape hatch set, matching the existing CLAUDECODE tests. Agentflare-Agent: claude-code_2-1-281_harness Agentflare-Branch: task/637-flare-git-shim-agent-settable-agentflare Agentflare-Item: 637 Agentflare-Session: 85157ad7-4af8-4643-ae34-bcbbf05e7434
getappz
added a commit
that referenced
this pull request
Sep 24, 2026
…ical-mutate guard (#803) * fix(git-shim): agent-invoked calls can no longer self-clear the canonical-mutate guard (item 637) deny_canonical_detach_reason checked AGENTFLARE_GIT_ALLOW_CANONICAL_MUTATE before checking agent_invocation_detected(), so an agent denied by the worktree-mandatory guard could set the override itself and mutate the canonical checkout directly -- this happened twice in practice, most recently detaching the user's main working-directory HEAD. Reorder the checks so the override is only reachable for non-agent-invoked (human/interactive) calls; the stranded-checkout checkout/switch recovery path (main.rs:502) is left untouched since it's narrow and already audited. Flip the existing escape-hatch tests to assert denial for agent-invoked callers and add a test confirming human/interactive use is unaffected. Agentflare-Agent: claude-code_2-1-281_harness Agentflare-Branch: task/637-flare-git-shim-agent-settable-agentflare Agentflare-Item: 637 Agentflare-Session: e402be06-2dbb-4058-bb61-bf38039c7d27 * fix(git-shim): classify LEAN_CTX_AGENT as agent-invoked (item #637 follow-up) CodeRabbit on #803 asked for a test exercising the dispatcher route rather than the compiled git shim directly. agentflare-shim's PATH dispatcher sets LEAN_CTX_AGENT before routing a command through lean-ctx -c git, but that marker was in neither agent-detector 0.2.1's catalog nor the AGENTFLARE_AGENT check, so a caller reaching this shim only via that route would misclassify as human and bypass the canonical-mutate guard entirely. - agent_invocation_detected() now also checks LEAN_CTX_AGENT directly. - human_shim strips it too, since it is ambient under lean-ctx sessions. - New test proves both branch-create and HEAD-detach stay denied under this marker, with the escape hatch set, matching the existing CLAUDECODE tests. Agentflare-Agent: claude-code_2-1-281_harness Agentflare-Branch: task/637-flare-git-shim-agent-settable-agentflare Agentflare-Item: 637 Agentflare-Session: 85157ad7-4af8-4643-ae34-bcbbf05e7434
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
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.
Only
repo-compare.jsonusesrun_if, so this change is isolated. Done.What I did: Extended
compile_run_ifincrates/flare-workflow/src/json.rsto support a single top-levelOR/ANDcombining two atomic conditions (each still==/!=/bare-truthy), and fixed a landmine I hit along the way:expand_variablesleaves an unresolved{{var}}as literal template text rather than empty string, which would have madeverdict != ''spuriously true whenevercomparenever ran. I now normalize any still-bracketed{{...}}output to falsy/empty insiderun_ifevaluation only (prompt expansion elsewhere is untouched, preserving the "agent sees the raw placeholder and refuses to fabricate" fallback).Wired
record's gate in.agentflare/workflows/repo-compare.jsonto{{cache_check}} != 'MISS' OR {{verdict}} != ''— skips only on a fresh-analysis run with no verdict, still runs on a legitimate cache-hit passthrough.Tests: added two unit tests directly against
compile_run_if(run_if_or_treats_unset_var_as_falsy_not_its_own_placeholder_text, covering MISS+no-verdict skip / MISS+verdict run / cache-hit run; andrun_if_and_requires_both_sides). Fullflare-workflowcrate suite: 29/29 passing, no regressions. JSON file validated as well-formed. Confirmed no other workflow file in the repo usesrun_if, so this grammar change is isolated torepo-compare.json.Concerns: the
OR/ANDsplit is a plain substring match, so a quoted literal containing the literal substring" OR "or" AND "would misparse — acceptable per the task's own scope note ("string-only, no full expression language"), but worth remembering if this grammar gets reused elsewhere.Opened by
claude-codeon flared:51bb8de6c33b for item #181 via agentflare.Summary by CodeRabbit
ORandANDlogic for more flexible automation.