diff options
Diffstat (limited to 'internal/executor')
| -rw-r--r-- | internal/executor/executor.go | 27 | ||||
| -rw-r--r-- | internal/executor/summary.go | 1 |
2 files changed, 20 insertions, 8 deletions
diff --git a/internal/executor/executor.go b/internal/executor/executor.go index 8614481..2f01b93 100644 --- a/internal/executor/executor.go +++ b/internal/executor/executor.go @@ -505,7 +505,7 @@ func (p *Pool) handleRunResult(ctx context.Context, t *task.Task, exec *storage. if err := p.store.UpdateTaskState(t.ID, task.StateReady); err != nil { p.logger.Error("failed to update task state", "taskID", t.ID, "state", task.StateReady, "error", err) } - go p.spawnCheckerTask(context.Background(), t) + go p.spawnCheckerTask(context.Background(), t, exec) } } else { exec.Status = "COMPLETED" @@ -629,7 +629,7 @@ func (p *Pool) ShipStory(ctx context.Context, storyID string) error { // spawnCheckerTask creates and submits a checker task for the given completed task. // Guards: not called for subtasks, checker tasks, tasks without a repository URL, // or tasks that already have a checker. -func (p *Pool) spawnCheckerTask(ctx context.Context, checked *task.Task) { +func (p *Pool) spawnCheckerTask(ctx context.Context, checked *task.Task, exec *storage.Execution) { // Never spawn a checker for subtasks, checker tasks, or tasks without a repository. if checked.ParentTaskID != "" || checked.CheckerForTaskID != "" || checked.RepositoryURL == "" { return @@ -649,6 +649,15 @@ func (p *Pool) spawnCheckerTask(ctx context.Context, checked *task.Task) { criteria = checked.Agent.Instructions } + // Embed the execution log so the checker can verify output without needing + // Docker, API access, or repository cloning. + logSnippet := "" + if exec != nil && exec.StdoutPath != "" { + logSnippet = "\n\nExecution log (last 200 lines):\n<execution_log>\n" + + tailFile(exec.StdoutPath, 200) + + "\n</execution_log>" + } + instructions := fmt.Sprintf(`You are validating a completed task. Do not make any changes to the code or repository. Task: %s @@ -656,15 +665,17 @@ Instructions given to the implementor: %s Acceptance criteria: -%s +%s%s Steps: -1. Clone the repository and review the changes made. -2. Verify each acceptance criterion is met. Run tests or make HTTP requests as needed. -3. If all criteria are satisfied, exit normally (success). -4. If any criterion is not met, use the Bash tool to exit with a non-zero code: +1. Review the execution log above to determine what the agent did. +2. If the task involved code changes, use Read/Grep/Glob to inspect the repository at %s. +3. Verify each acceptance criterion is met. Run tests or make HTTP requests as needed. +4. If all criteria are satisfied, exit normally (success). +5. If any criterion is not met, use the Bash tool to exit with a non-zero code: bash -c "exit 1" - Before exiting, write a brief summary of what failed.`, checked.Name, checked.Agent.Instructions, criteria) + Before exiting, write a brief summary of what failed.`, + checked.Name, checked.Agent.Instructions, criteria, logSnippet, checked.RepositoryURL) now := time.Now().UTC() checker := &task.Task{ diff --git a/internal/executor/summary.go b/internal/executor/summary.go index bcf5cfd..e5a1772 100644 --- a/internal/executor/summary.go +++ b/internal/executor/summary.go @@ -12,6 +12,7 @@ import ( "github.com/thepeterstone/claudomator/internal/llm" ) + // synthesizeSummaryMaxBytes caps how much of the stdout log we send to the // LLM. Larger values cost more tokens with diminishing returns for a 2-4 // sentence summary. |