Sid c9f0d6ac8e feat(agents): support thinkingDefault: "adaptive" for Anthropic models (#31227) ...

* feat(agents): support `thinkingDefault: "adaptive"` for Anthropic models

Anthropic's Opus 4.6 and Sonnet 4.6 support adaptive thinking where the
model dynamically decides when and how much to think.  This is now
Anthropic's recommended mode and `budget_tokens` is deprecated on these
models.

Add "adaptive" as a valid thinking level:
- Config: `agents.defaults.thinkingDefault: "adaptive"`
- CLI: `/think adaptive` or `/think auto`
- Pi SDK mapping: "adaptive" → "medium" effort at the pi-agent-core
  layer, which the Anthropic provider translates to
  `thinking.type: "adaptive"` with `output_config.effort: "medium"`
- Provider fallbacks: OpenRouter and Google map "adaptive" to their
  respective "medium" equivalents

Closes #30880

Made-with: Cursor

* style(changelog): format changelog with oxfmt

* test(types): fix strict typing in runtime/plugin-context tests

---------

Co-authored-by: Peter Steinberger <steipete@gmail.com>

2026-03-02 03:52:02 +00:00

..

run

fix(agents): unblock gpt-5.3-codex API-key routing and replay (#31083)

2026-03-02 03:45:12 +00:00

abort.ts

Update contributing, deduplicate more functions

2026-02-09 19:21:33 -08:00

cache-ttl.test.ts

fix(agents): extend cache-ttl eligibility for moonshot and zai

2026-02-23 18:27:36 +00:00

cache-ttl.ts

feat: Add Kilo Gateway provider (#20212)

2026-02-23 23:29:27 +00:00

compact.ts

fix(agents): harden embedded pi project settings loading

2026-02-26 21:46:39 +01:00

compaction-safety-timeout.ts

fix: add safety timeout to session.compact() to prevent lane deadlock (#16533)

2026-02-14 17:54:12 -05:00

extensions.ts

Compaction: preserve opaque identifiers in summaries (openclaw#25553) thanks @rodrigouroz

2026-02-27 08:14:05 -06:00

extra-params.cache-retention-default.test.ts

feat: surface cached token counts in /status output (openclaw#21248) thanks @vishaltandale00

2026-02-19 21:06:13 -06:00

extra-params.openrouter-cache-control.test.ts

test: dedupe repeated test fixtures and assertions

2026-02-22 18:37:25 +00:00

extra-params.ts

feat(agents): support thinkingDefault: "adaptive" for Anthropic models (#31227)

2026-03-02 03:52:02 +00:00

extra-params.zai-tool-stream.test.ts

docs(zai): document tool_stream defaults

2026-02-17 09:22:55 -05:00

google.test.ts

refactor!: remove google-antigravity provider support

2026-02-23 05:20:14 +01:00

google.ts

fix(agents): unblock gpt-5.3-codex API-key routing and replay (#31083)

2026-03-02 03:45:12 +00:00

history.ts

fix(discord): Apply historyLimit to channel/group sessions to prevent compaction bypass (openclaw#11356) thanks @shadril238

2026-02-13 18:18:57 -06:00

kilocode.test.ts

feat: Add Kilo Gateway provider (#20212)

2026-02-23 23:29:27 +00:00

lanes.ts

refactor: use command lane enum

2026-01-20 10:51:25 +00:00

logger.ts

refactor(logging): split config + subsystem imports

2026-01-19 00:15:44 +00:00

model.forward-compat.test.ts

fix(agents): add forward-compat fallback for google-gemini-cli gemini-3.1-pro/flash-preview (#26570)

2026-02-26 18:39:13 -05:00

model.test-harness.ts

fix(agents): add forward-compat fallback for google-gemini-cli gemini-3.1-pro/flash-preview (#26570)

2026-02-26 18:39:13 -05:00

model.test.ts

fix(model): preserve reasoning in provider fallback resolution (#29285)

2026-02-27 17:38:22 -08:00

model.ts

fix(model): preserve reasoning in provider fallback resolution (#29285)

2026-02-27 17:38:22 -08:00

run.overflow-compaction.fixture.ts

refactor(test): dedupe agent and discord test fixtures

2026-02-22 20:04:51 +00:00

run.overflow-compaction.loop.test.ts

refactor(test): dedupe agent and discord test fixtures

2026-02-22 20:04:51 +00:00

run.overflow-compaction.mocks.shared.ts

test: dedupe fixtures and test harness setup

2026-02-23 05:45:54 +00:00

run.overflow-compaction.shared-test.ts

test(pi): share overflow-compaction test setup

2026-02-22 07:44:57 +00:00

run.overflow-compaction.test.ts

refactor(test): dedupe agent and discord test fixtures

2026-02-22 20:04:51 +00:00

run.ts

fix: refresh Copilot token before expiry and retry on auth errors

2026-03-02 02:09:40 +00:00

runs.ts

fix(discord): preserve channel session keys via channel_id fallbacks (#17622)

2026-02-15 20:30:17 -06:00

sandbox-info.ts

fix (sandbox/prompts): align workspace guidance with container workdir

2026-02-14 20:20:42 -08:00

sanitize-session-history.tool-result-details.test.ts

test: move embedded and tool agent suites out of e2e

2026-02-22 11:17:47 +00:00

session-manager-cache.ts

chore: Enable "experimentalSortImports" in Oxfmt and reformat all imorts.

2026-02-01 10:03:47 +09:00

session-manager-init.ts

revert: PR 18288 accidental merge (#19224)

2026-02-17 10:05:29 -05:00

system-prompt.test.ts

fix(plugins): hook systemPrompt gets collected then thrown away (#14583) (#14602)

2026-02-22 18:58:21 -05:00

system-prompt.ts

feat: ACP thread-bound agents (#23580)

2026-02-26 11:00:09 +01:00

thinking.test.ts

refactor(channels): dedupe hook and monitor execution paths

2026-02-22 21:19:09 +00:00

thinking.ts

refactor(channels): dedupe hook and monitor execution paths

2026-02-22 21:19:09 +00:00

tool-name-allowlist.ts

Agents: validate persisted tool-call names

2026-02-21 23:06:44 -08:00

tool-result-context-guard.test.ts

refactor(channels): dedupe hook and monitor execution paths

2026-02-22 21:19:09 +00:00

tool-result-context-guard.ts

fix(subagent): harden read-tool overflow guards and sticky reply threading (#19508)

2026-02-17 15:32:52 -08:00

tool-result-truncation.test.ts

refactor(agents): dedupe config and truncation guards

2026-02-22 17:54:51 +00:00

tool-result-truncation.ts

refactor(agents): dedupe config and truncation guards

2026-02-22 17:54:51 +00:00

tool-split.ts

chore: Enable "experimentalSortImports" in Oxfmt and reformat all imorts.

2026-02-01 10:03:47 +09:00

types.ts

fix(agents): cap embedded runner retry loop

2026-02-21 15:35:45 +01:00

usage-reporting.test.ts

chore: Fix types in tests 43/N.

2026-02-17 15:50:07 +09:00

utils.ts

feat(agents): support thinkingDefault: "adaptive" for Anthropic models (#31227)

2026-03-02 03:52:02 +00:00

wait-for-idle-before-flush.ts

fix(agents): wait for agent idle before flushing pending tool results (#13746)

2026-02-13 20:35:43 +01:00