Capability map¶
This map is the task-oriented index to the complete documentation set. The curated navigation keeps the common path short; this page makes every implemented capability and design record reachable without knowing its name in advance. For exact signatures, use the generated API reference. For executable code, use the runnable cookbook.
Agent authoring and execution¶
- Agent identity — bind a run to the caller it acts for.
- Activity policies — bound retries, timeouts, and activity behavior.
- Audit — record security- and operation-relevant decisions.
- Autonomy — declare how independently an agent may act.
- Delegation and peer delegation — narrow authority across child and remote work.
- Dispatch — derive an execution schedule from declared dependencies.
- Estimation — estimate a run before admitting live spend.
- Execution models — choose in-process, queued, or durable execution.
- Parallel fan-out — aggregate concurrent branches without hiding failures.
- Planning — separate model-authored plans from deterministic execution.
- Suspension — persist approval waits and resume them safely.
- Compensation — account for effects that need deterministic reversal.
- Escalation ladder — choose the smallest agent architecture that fits the task.
Prompts, context, and outputs¶
- Context assembly and conversation compaction.
- Content-typed compression, reversible compression, and the frozen prefix.
- Output validation and output shaping.
- Prompt registry, templates, linting, evaluation gate, and rollback.
- Response caching and token-optimizer routing.
- Savings accounting — distinguish measured input savings from estimated output savings.
Knowledge, memory, and retrieval¶
- Memory admission and provenance and graph memory.
- Document and audio intake.
- Chunking, embedding, embedding batching, and local embedding and reranking.
- Reranking, citations, and corpus-poisoning defenses.
Tools, code, and remote boundaries¶
- Tool approval, concurrency, credentials, and typed errors.
- Code intelligence and sandboxed code execution.
- Credential rotation and secret resolution.
- PII redaction and content policy.
- MCP authentication context, credentials, resilience, server export, and surface pinning.
State, events, and recovery¶
- Event contracts and JetStream delivery.
- Spend ledger and versioned persisted state.
- Durable runs, checkpointing, transactional outbox, and dead-letter replay.
Testing, evaluation, and observability¶
- Determinism and replay, provider cassettes, and the scripted fake provider.
- Benchmarks, compression gates, and the calibrated evaluation judge.
- Latency objectives, local trace view, and spend and performance metrics.
Developer workflows¶
- Command-line guide for the installed
tesserix-adkcommand. - Runnable cookbook for offline examples mapped to every public symbol.
- Testing guide, reference agents, and migration guide.