Batch Daily Go Test Parallelizer to process up to 5 files per run

.github/skills/agentic-workflows/SKILL.md

@@ -71,6 +71,7 @@ Load these files from `github/gh-aw` (they are not available locally).
 - `.github/aw/test-coverage.md`
 - `.github/aw/test-expression.md`
 - `.github/aw/token-optimization-caching-budgets.md`
+- `.github/aw/token-optimization-observability.md`
 - `.github/aw/token-optimization.md`
 - `.github/aw/triggers.md`
 - `.github/aw/update-agentic-workflow.md`

.github/workflows/daily-go-test-parallelizer.md

@@ -35,7 +35,7 @@ safe-outputs:
     protected-files: blocked
     allowed-files:
       - "**/*_test.go"
-    max-patch-files: 1
+    max-patch-files: 25
     max-patch-size: 2048
   noop:
 evals:
@@ -47,20 +47,29 @@ evals:
 
 # Daily Go Test Parallelizer
 
-Analyze one Go test file per run and add `t.Parallel()` only where parallel execution is demonstrably safe.
+Analyze up to twenty-five Go test files per run and add `t.Parallel()` only where parallel execution is demonstrably safe.
 
 ## Select a file
 
 1. Use `grep` to list tracked `*_test.go` files containing top-level `Test` functions. Exclude `vendor/` and generated files, then sort paths lexicographically.
 2. Read `/tmp/gh-aw/cache-memory/go-test-parallelizer/state.json` when it exists. It has this shape:
    `{"last_file":"path/to/file_test.go"}`.
-3. Select the path after `last_file`, wrapping to the first path. If the cache is absent, malformed, or names a removed file, select the first path.
-4. Analyze and modify at most that one file.
+3. Select up to 25 consecutive paths starting after `last_file`, wrapping to the first path. If the cache is absent, malformed, or names a removed file, start from the first path.
+4. Analyze and modify only this selected batch (maximum 25 files).
+5. Minimize token usage: avoid re-reading files you already analyzed, and do not paste full file contents into notes or outputs.
 
 ## Analyze safety
 
 Add `t.Parallel()` at the start of eligible top-level tests. Also add it to eligible table-driven subtests when loop variables are safely captured.
 
+## Inline analysis agents (small context)
+
+1. For each selected path, call `parallel-safety-checker` exactly once with only that file path and the safety rules from this workflow.
+2. Dispatch these inline agent calls in one parallel tool-use block when possible. If not possible, dispatch in small batches of 5.
+3. Require compact JSON output from each sub-agent:
+   `{"file":"...","safe":true|false,"reasons":["..."],"candidate_tests":["TestName"]}`.
+4. Use the JSON results to decide which files to edit. Keep aggregation notes short and avoid repeating unchanged rule text.
+
 Do not parallelize tests that use or may conflict through:
 
 - `t.Setenv`, `os.Setenv`, `os.Chdir`, or other process-wide state
@@ -72,15 +81,25 @@ Do not change assertions, test behavior, production code, dependencies, generate
 
 ## Validate
 
-After editing:
+After editing the selected batch:
 
-1. Run the selected package with the race detector.
-2. Run `go test ./...`.
-3. Inspect the diff and confirm it contains only safe `t.Parallel()` additions in the selected file.
-4. Revert the edit and use `noop` if either test command fails or the diff contains any other change.
+1. Run `go test -race` once per unique modified package (deduplicate package paths across edited files).
+2. Run `go test ./...` once after all candidate edits.
+3. Inspect the diff and confirm it contains only safe `t.Parallel()` additions in selected files.
+4. Revert the edit and use `noop` if any test command fails or the diff contains any other change.
 
 ## Persist and report
 
-Always create `/tmp/gh-aw/cache-memory/go-test-parallelizer/` and write the selected path to `state.json`, even when no edit is safe, so the next daily run advances round-robin.
+Always create `/tmp/gh-aw/cache-memory/go-test-parallelizer/` and write the last path from the selected batch to `state.json`, even when no edit is safe, so the next daily run advances round-robin.
 
 If validation succeeds with a change, create one draft pull request describing the safety analysis and test results. Otherwise use `noop` with the selected path and a short reason.
+
+## agent: `parallel-safety-checker`
+---
+description: Review one Go test file for safe t.Parallel additions with minimal context
+model: gpt-5-mini
+---
+Given one `*_test.go` file path, read only that file and apply this workflow's safety rules.
+Return compact JSON only in this exact shape:
+`{"file":"...","safe":true|false,"reasons":["..."],"candidate_tests":["TestName"]}`.
+Set `safe` to false when uncertain.