feat(observability): LLM-assisted trace analysis for failed tasks

**Context:** ROADMAP.md → Agent quality → LLM-assisted trace analysis
**Related:** `--trace` (shipped), automated alert triage draft

---

## Component

API or orchestration

## Describe the feature

Automated **post-mortem** deep dive on failed task trajectories (S3 trace NDJSON + OpenTelemetry spans + `TaskEvents`) using an LLM to surface recurring reasoning and tool-use failure modes. Distinct from mid-run `bgagent ask` and optional status summarization.

## Use case

Operators debugging failure clusters spend hours reading traces. Platform teams need pattern detection across failures (e.g. repeated `git push` auth errors, hallucinated file paths).

## Proposed solution

1. Trigger on `task_failed` (optional Blueprint flag or operator-initiated `bgagent trace analyze`).
2. Lambda loads trace + events; calls Bedrock with structured output schema.
3. Persist analysis artifact (S3 + `TaskEvents` `trace_analysis` event).
4. Rate limits and cost caps per analysis.
5. Optional aggregation job for cross-task themes (weekly operator report).

## Other information

- Distinct from **automated alert triage** (security/ops alarms, not task post-mortems).
- Design context: `docs/design/OBSERVABILITY.md`, `docs/design/EVALUATION.md`.

- [ ] This might be a breaking change

Inizia da ROADMAP.md, docs/design/OBSERVABILITY.md e docs/design/EVALUATION.md, quindi traccia il flusso --trace esistente e il percorso dell’evento task_failed. Esamina l’entry point proposto bgagent trace analyze e definisci come si integrano i dati di trace, l’analisi Bedrock, gli artefatti persistiti, i limiti di frequenza e l’aggregazione opzionale. Il lavoro è completato quando l’ambito e il design della funzionalità sono definiti per l’implementazione.

Scritto dal modello di indicizzazione a partire dal testo della issue.