Skip to content

feat(aux): translate extra_body.reasoning into Codex Responses API#17004

Merged
teknium1 merged 1 commit into
mainfrom
hermes/hermes-58b9453f
Apr 28, 2026
Merged

feat(aux): translate extra_body.reasoning into Codex Responses API#17004
teknium1 merged 1 commit into
mainfrom
hermes/hermes-58b9453f

Conversation

@teknium1

@teknium1 teknium1 commented Apr 28, 2026

Copy link
Copy Markdown
Contributor

Summary

Auxiliary callers can now steer Codex Responses API reasoning via auxiliary.<task>.extra_body.reasoning — previously this was silently dropped by _CodexCompletionsAdapter.

Root cause

The adapter's .create() only forwarded messages / model / tools into responses.stream(). Any extra_body.reasoning hint that worked fine against chat-completions providers disappeared for Codex aux calls.

Changes

  • agent/auxiliary_client.py: _CodexCompletionsAdapter.create() reads extra_body.reasoning and translates it into the Responses API's top-level reasoning + include fields, mirroring agent/transports/codex.py::build_kwargs()
  • tests/agent/test_auxiliary_client.py: new TestCodexAdapterReasoningTranslation class with 9 tests

Translation rules (match main-agent Codex transport)

extra_body.reasoning resp_kwargs
{"effort": "medium"} reasoning={effort:medium, summary:auto}, include=[reasoning.encrypted_content]
{"effort": "minimal"} clamped to low
{"effort": "high"} passed through
{"enabled": False} reasoning + include omitted
{} (truthy) defaults to medium
missing / non-dict no-op

Validation

scripts/run_tests.sh tests/agent/test_auxiliary_client.py::TestCodexAdapterReasoningTranslation — 9 passed. Full test_auxiliary_client.py — 105 passed, 1 pre-existing failure unrelated to this change (verified against origin/main before patch).

Credit

Reported by @op in the Apr 26 feedback bundle.

@Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API
e3d0f18 
Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs
@teknium1 teknium1 merged commit 391f1ca into main Apr 28, 2026
11 of 12 checks passed
@teknium1 teknium1 deleted the hermes/hermes-58b9453f branch April 28, 2026 12:47
@alt-glitch alt-glitch added type/feature New feature or request P2 Medium — degraded but workaround exists comp/agent Core agent loop, run_agent.py, prompt builder provider/openai OpenAI / Codex Responses API labels Apr 28, 2026
@briandevans briandevans mentioned this pull request Apr 28, 2026
Closed
4 tasks
cluricaun28 referenced this pull request in cluricaun28/Logos Apr 28, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (#…
b90a371 
…17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
ulasbilgen pushed a commit to ulasbilgen/hermes-adhd-agent that referenced this pull request May 1, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
f6187f7 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
@github-actions github-actions Bot mentioned this pull request May 1, 2026
Merged
donald131 pushed a commit to donald131/hermes-agent that referenced this pull request May 2, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
26831ef 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
@voteblake voteblake mentioned this pull request May 7, 2026
Closed
1 task
02356abc pushed a commit to 02356abc/hermes-agent that referenced this pull request May 14, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
35191d4 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
dannyJ848 pushed a commit to dannyJ848/hermes-agent that referenced this pull request May 17, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
973e71b 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
aashizpoudel pushed a commit to aashizpoudel/hermes-agent that referenced this pull request May 21, 2026
@teknium1 @Teknium @aashizpoudel
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
a4c8dac 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
gweeteve pushed a commit to gweeteve/hermes-agent that referenced this pull request Jun 2, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
6460ef1 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
dev-xyz-0-0 added a commit to dev-xyz-0-0/hermes-agent that referenced this pull request Jun 7, 2026
@dev-xyz-0-0
1. Codex auxiliary calls now honor reasoning settings.
3939e72 
2. Main-agent and auxiliary Codex behavior are consistent.
3. "minimal" effort is safely clamped to "low" to avoid backend rejection.
4. Badly shaped reasoning config is ignored safely instead of crashing.
5. Tests cover the full translation matrix.  NousResearch/hermes-agent#17004
Egavasyug pushed a commit to Egavasyug/hermes-agent that referenced this pull request Jun 10, 2026
@teknium1 @Teknium
feat(aux): translate extra_body.reasoning into Codex Responses API (N…
c05828c 
…ousResearch#17004)

Auxiliary callers that configure reasoning via
auxiliary.<task>.extra_body.reasoning were having that config silently
dropped by the Codex Responses adapter — it only forwarded
messages/model/tools through to responses.stream(), never translating
chat.completions-shaped reasoning hints into the Responses API's
top-level reasoning + include fields.

Mirror the main-agent translation from agent/transports/codex.py:
- extra_body.reasoning.effort → resp_kwargs.reasoning.{effort, summary:"auto"}
- 'minimal' → 'low' clamp (Codex backend rejects 'minimal')
- Always include ['reasoning.encrypted_content'] when reasoning is enabled
- {'enabled': False} → omit reasoning and include entirely
- Non-dict reasoning values are ignored defensively

Reported by @op (Apr 26 feedback bundle).

## Changes
- agent/auxiliary_client.py: _CodexCompletionsAdapter.create() now reads
  and translates extra_body.reasoning before calling responses.stream()
- tests/agent/test_auxiliary_client.py: 9 new tests covering all effort
  levels, the minimal→low clamp, the disabled path, the no-op paths,
  and defensive handling of wrong-shape inputs

Co-authored-by: teknium1 <teknium@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

comp/agent Core agent loop, run_agent.py, prompt builder P2 Medium — degraded but workaround exists provider/openai OpenAI / Codex Responses API type/feature New feature or request

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@teknium1 @alt-glitch @Teknium