Skip to content

feat(cli): add [INFO] log prefix, fix verdict icons, show score %#1074

Merged
christso merged 8 commits intomainfrom
feat/1073-agentv-log-prefix
Apr 13, 2026
Merged

feat(cli): add [INFO] log prefix, fix verdict icons, show score %#1074
christso merged 8 commits intomainfrom
feat/1073-agentv-log-prefix

Conversation

@christso
Copy link
Copy Markdown
Collaborator

@christso christso commented Apr 13, 2026

Summary

Closes #1073

  • Fix verdict icons: completed tests with FAIL verdict now show ⚠️ instead of ✅, and ERROR shows ❌
  • Show scores as percentages (75%) instead of decimals (0.750)
  • Show resolved target name when default is a use_target redirect (e.g., default → copilot)
  • Replace grouped "Copilot CLI logs:" heading with per-line Provider log: prefix that stays meaningful when interleaved

Before

Using target: default
0/11   🔄 queue-no-requirements | default
1/11   ✅ some-failing-test | default | 0.500 PASS

Copilot CLI logs:
1. /path/to/log1
2. /path/to/log2

After

Using target: default → copilot
0/11   🔄 queue-no-requirements | default → copilot
1/11   ⚠️ some-failing-test | default → copilot | 50% FAIL
2/11   ❌ broken-test | default → copilot | ERROR
Provider log: /path/to/log1
Provider log: /path/to/log2

Test plan

  • All 2123 tests pass (1613 core + 67 eval SDK + 443 CLI)
  • Build, typecheck, lint all pass
  • Manual UAT: dry-run eval confirms icons (✅/⚠️/❌), percentage scores, resolved target names, and clean summary block

🤖 Generated with Claude Code

christso and others added 2 commits April 13, 2026 05:35
…rcentage

Add [INFO] prefix to all eval CLI output lines so framework messages are
distinguishable from interleaved provider logs (Copilot, Codex, Pi).

Fix verdict icons: completed tests with FAIL verdict now show ❌ instead of ✅,
and ERROR shows ⚠️. Show scores as percentages (75%) instead of decimals (0.750).

Closes #1073

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@cloudflare-workers-and-pages
Copy link
Copy Markdown

cloudflare-workers-and-pages bot commented Apr 13, 2026

Deploying agentv with  Cloudflare Pages  Cloudflare Pages

Latest commit: 3c43848
Status: ✅  Deploy successful!
Preview URL: https://81d8e4d3.agentv.pages.dev
Branch Preview URL: https://feat-1073-agentv-log-prefix.agentv.pages.dev

View logs

christso and others added 6 commits April 13, 2026 05:39
FAIL is a quality warning (test ran but scored below threshold).
ERROR is a hard failure (execution broke). Icons now match severity.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The evaluation summary and matrix table are already delimited with
===== borders and print as a single block after all tests complete.
The prefix is only needed for progress lines interleaved with provider
output mid-run.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The "Copilot CLI logs:" heading lost context when interleaved with
provider output. Each log path now prints as its own self-contained
line: [INFO] Log created: /path/to/log

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…irect

When "default" in targets.yaml delegates via use_target (e.g., to
"copilot"), the progress lines now show "default → copilot" instead
of just "default", so the user sees which provider is actually running.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ient

With verdict icons (✅/⚠️/❌), percentages, and resolved target names
(default → copilot), every line is already self-describing. The [INFO]
prefix was just noise.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@christso christso marked this pull request as ready for review April 13, 2026 06:36
@christso christso merged commit 4506eb1 into main Apr 13, 2026
4 checks passed
@christso christso deleted the feat/1073-agentv-log-prefix branch April 13, 2026 06:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

feat(cli): Add [AGENTV] log prefix, fix verdict icons, show score as percentage

1 participant