agentv 2.9.0-next.1 → 2.10.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (22) hide show
  1. package/dist/{chunk-H54JIK7G.js → chunk-G3OTPFYX.js} +2 -3
  2. package/dist/chunk-G3OTPFYX.js.map +1 -0
  3. package/dist/cli.js +1 -1
  4. package/dist/index.js +1 -1
  5. package/dist/templates/.agentv/config.yaml +1 -1
  6. package/dist/templates/.agentv/targets.yaml +10 -13
  7. package/package.json +1 -1
  8. package/dist/chunk-H54JIK7G.js.map +0 -1
  9. package/dist/templates/.claude/skills/agentv-eval-builder/SKILL.md +0 -202
  10. package/dist/templates/.claude/skills/agentv-eval-builder/references/batch-cli-evaluator.md +0 -316
  11. package/dist/templates/.claude/skills/agentv-eval-builder/references/compare-command.md +0 -137
  12. package/dist/templates/.claude/skills/agentv-eval-builder/references/composite-evaluator.md +0 -215
  13. package/dist/templates/.claude/skills/agentv-eval-builder/references/config-schema.json +0 -27
  14. package/dist/templates/.claude/skills/agentv-eval-builder/references/custom-evaluators.md +0 -118
  15. package/dist/templates/.claude/skills/agentv-eval-builder/references/eval-schema.json +0 -278
  16. package/dist/templates/.claude/skills/agentv-eval-builder/references/example-evals.md +0 -333
  17. package/dist/templates/.claude/skills/agentv-eval-builder/references/rubric-evaluator.md +0 -77
  18. package/dist/templates/.claude/skills/agentv-eval-builder/references/structured-data-evaluators.md +0 -121
  19. package/dist/templates/.claude/skills/agentv-eval-builder/references/tool-trajectory-evaluator.md +0 -298
  20. package/dist/templates/.claude/skills/agentv-prompt-optimizer/SKILL.md +0 -78
  21. package/dist/templates/.github/prompts/agentv-eval-build.prompt.md +0 -5
  22. package/dist/templates/.github/prompts/agentv-optimize.prompt.md +0 -4
package/dist/cli.js CHANGED
@@ -2,7 +2,7 @@
2
2
  import { createRequire } from 'node:module'; const require = createRequire(import.meta.url);
3
3
  import {
4
4
  runCli
5
- } from "./chunk-H54JIK7G.js";
5
+ } from "./chunk-G3OTPFYX.js";
6
6
  import "./chunk-PC3FAOHT.js";
7
7
  import "./chunk-RJWTL3VS.js";
8
8
  import "./chunk-JK6V4KVD.js";
package/dist/index.js CHANGED
@@ -3,7 +3,7 @@ import {
3
3
  app,
4
4
  preprocessArgv,
5
5
  runCli
6
- } from "./chunk-H54JIK7G.js";
6
+ } from "./chunk-G3OTPFYX.js";
7
7
  import "./chunk-PC3FAOHT.js";
8
8
  import "./chunk-RJWTL3VS.js";
9
9
  import "./chunk-JK6V4KVD.js";
@@ -12,4 +12,4 @@ guideline_patterns:
12
12
  # - Patterns use standard glob syntax (via micromatch library)
13
13
  # - Paths are normalized to forward slashes for cross-platform compatibility
14
14
  # - Only files matching these patterns are loaded as guidelines
15
- # - All other files referenced in eval cases are treated as regular file content
15
+ # - All other files referenced in tests are treated as regular file content
@@ -31,19 +31,17 @@ targets:
31
31
  log_dir: ${{ CODEX_LOG_DIR }} # Optional: where Codex CLI stream logs are stored (defaults to ./.agentv/logs/codex)
32
32
  log_format: json # Optional: 'summary' (default) or 'json' for raw event logs
33
33
 
34
- # Claude Code - Anthropic's autonomous coding CLI
35
- - name: claude-code
36
- provider: claude-code
34
+ # Claude - Anthropic's Claude Agent SDK
35
+ - name: claude
36
+ provider: claude
37
37
  judge_target: azure_base
38
- # Uses the Claude Code CLI (defaults to `claude` on PATH)
39
- # executable: ${{ CLAUDE_CODE_CLI_PATH }} # Optional: override executable path
38
+ # Uses the @anthropic-ai/claude-agent-sdk
40
39
  # model: claude-sonnet-4-20250514 # Optional: override model
41
- # args: # Optional additional CLI arguments
42
- # - --allowedTools
43
- # - Read,Write,Edit,Bash
44
40
  timeout_seconds: 180
45
- # cwd: ${{ CLAUDE_CODE_WORKSPACE_DIR }} # Optional: working directory (defaults to process.cwd())
46
- # log_dir: ${{ CLAUDE_CODE_LOG_DIR }} # Optional: where stream logs are stored (defaults to ./.agentv/logs/claude-code)
41
+ # cwd: ${{ CLAUDE_WORKSPACE_DIR }} # Optional: working directory (defaults to process.cwd())
42
+ # max_turns: 50 # Optional: max conversation turns
43
+ # max_budget_usd: 5.0 # Optional: max cost budget in USD
44
+ # log_dir: ${{ CLAUDE_LOG_DIR }} # Optional: where stream logs are stored (defaults to ./.agentv/logs/claude)
47
45
  log_format: json # Optional: 'summary' (default) or 'json' for raw event logs
48
46
  # system_prompt: optional override (default instructs agent to include code in response)
49
47
 
@@ -76,12 +74,11 @@ targets:
76
74
  judge_target: azure_base
77
75
  # Passes the fully rendered prompt and any attached files to a local Python script
78
76
  # NOTE: Do not add quotes around {PROMPT} or {FILES} - they are already shell-escaped
79
- command_template: uv run ./mock_cli.py --prompt {PROMPT} {FILES} --output {OUTPUT_FILE}
77
+ command: uv run ./mock_cli.py --prompt {PROMPT} {FILES} --output {OUTPUT_FILE}
80
78
  # Format for each file in {FILES}. {path} and {basename} are automatically shell-escaped, so no quotes needed
81
79
  files_format: --file {path}
82
80
  # Optional working directory resolved from .env
83
81
  cwd: ${{ CLI_EVALS_DIR }}
84
82
  timeout_seconds: 30
85
83
  healthcheck:
86
- type: command
87
- command_template: uv run ./mock_cli.py --healthcheck
84
+ command: uv run ./mock_cli.py --healthcheck
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "agentv",
3
- "version": "2.9.0-next.1",
3
+ "version": "2.10.0",
4
4
  "description": "CLI entry point for AgentV",
5
5
  "type": "module",
6
6
  "repository": {