feat(agentv-bench): add executor subagent for non-CLI targets#808
Merged
feat(agentv-bench): add executor subagent for non-CLI targets#808
Conversation
Add executor subagent that performs eval test cases directly when the
target is a non-CLI provider. Mirrors the grader pattern — one executor
per test case, all dispatched in parallel.
- Create agents/executor.md with workspace isolation guardrails
- Update SKILL.md: executor dispatch flow, subagent_mode_allowed docs,
targets.yaml location and ${{ ENV_VAR }} security note
- Update eval-yaml-spec.md: manifest-based opt-out, two invoke.json kinds
- Update run_tests.py docstring
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Deploying agentv with
|
| Latest commit: |
9c7726c
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://9adc4eaf.agentv.pages.dev |
| Branch Preview URL: | https://feat-797-executor-subagent-s.agentv.pages.dev |
Executor should have access to all tools, not a hardcoded subset. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Rename input.json fields to match the eval YAML schema and code grader SDK naming conventions: - input_text + input_messages → input (Message[]) - file_paths → input_files (string[]) Drop redundant input_text — derive it from input[0].content where needed. Remove unnecessary user_notes step from executor subagent. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…er payload The code grader SDK uses input_text and expected_output_text (via toSnakeCaseDeep). The question and reference_answer fields were redundant legacy names not present in the real orchestrator payload. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The executor works in the current workspace naturally. No need to restrict or parameterize the working directory. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…agents Clarify that the current workspace is the target workspace (multi-repo), not the eval repo. Warn if the user opened the wrong workspace since executor subagents won't have access to the agent's skills and repos. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…rkspace-dependent evals Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Single message: returns content directly (no role prefix). Multiple messages: prefixes each with @ROLE for least surprise. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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.
Summary
agents/executor.md— executor subagent that performs eval test cases (mirrors grader pattern)subagent_mode_alloweddocs,.agentv/targets.yamllocation,${{ ENV_VAR }}security noteSkill-side companion to #804 and #807.
Test plan
🤖 Generated with Claude Code