fix(tui): surface backend error as visible text when final_response is empty by teknium1 · Pull Request #21245 · NousResearch/hermes-agent

teknium1 · 2026-05-07T12:53:15Z

Summary

TUI / one-shot CLI now show Error: <detail> when the backend fails with no visible output, instead of emitting a blank turn.

Reported by @counterposition via #20873. That PR proposed a broader structured-failure refactor (~1000 LOC, new helper module, Ink runtime rows, 14 call-site rewrites). This is the minimal fix for the reported symptom.

Root cause: tui_gateway/server.py and the one-shot CLI path read result.get("final_response", "") and dropped the error field on the floor. The interactive CLI path at cli.py:9832 already handled this correctly — just not the other two surfaces.

Repro

hermes --tui --provider nous --model kimi-k2.6 (valid slug is moonshotai/kimi-k2.6). Session opens, prompts appear to succeed, responses are blank.

Changes

tui_gateway/server.py — when final_response is empty AND (failed or partial) AND error is set, emit Error: <detail> as the visible text. Status field already flagged as "error".
cli.py (one-shot mode) — same surfacing, writes to stderr so piped stdout stays clean for automation wrappers.
Two new regression tests covering the fix and asserting the empty-without-error path stays untouched.

What stays the same

final_response: None with no error flags → still renders empty (intentional — 'model produced nothing visible' is distinct from 'backend error').
"(empty)" sentinel → still owned by existing sentinel handlers (gateway/run.py:6480, run_agent.py:6716).
Interactive CLI path (cli.py:9832) — unchanged, already did this.

Validation

	Before	After
`tests/test_tui_gateway_server.py -k prompt_submit`	7 passing	9 passing (2 new)
Invalid slug via TUI	blank turn	`Error: HTTP 400: invalid model id ...`
Empty successful turn	empty	empty (unchanged)
`"(empty)"` sentinel	existing handler	existing handler (unchanged)

Closes #20873 (via minimal alternative).

@counterposition

…s empty When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR #20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

github-actions · 2026-05-07T12:54:29Z

🔎 Lint report: `fix/tui-surface-backend-errors` vs `origin/main`

ruff

Total: 0 on HEAD, 0 on base (➖ 0)

🆕 New issues: none

✅ Fixed issues: none

Unchanged: 0 pre-existing issues carried over.

ty (type checker)

Total: 7520 on HEAD, 7504 on base (🆕 +16)

🆕 New issues (15):

Rule	Count
`invalid-argument-type`	8
`unresolved-attribute`	3
`unsupported-operator`	3
`not-subscriptable`	1

First entries

tests/run_agent/test_provider_attribution_headers.py:129: [not-subscriptable] not-subscriptable: Cannot subscript object of type `int` with no `__getitem__` method
tests/run_agent/test_provider_attribution_headers.py:65: [unresolved-attribute] unresolved-attribute: Attribute `startswith` is not defined on `dict[str, str]` in union `Unknown | str | dict[str, str]`
tests/agent/test_codex_cloudflare_headers.py:181: [unsupported-operator] unsupported-operator: Operator `in` is not supported between objects of type `Literal["originator"]` and `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | ... omitted 3 union elements`
run_agent.py:2437: [invalid-argument-type] invalid-argument-type: Argument to function `get_model_context_length` is incorrect: Expected `str`, found `str | dict[str, str] | Any | ... omitted 3 union elements`
run_agent.py:6467: [invalid-argument-type] invalid-argument-type: Argument to function `build_anthropic_client` is incorrect: Expected `str`, found `str | dict[Unknown | str, Unknown | str | dict[str, str]] | Any | ... omitted 3 union elements`
run_agent.py:12331: [invalid-argument-type] invalid-argument-type: Argument to function `len` is incorrect: Expected `Sized`, found `(str & ~AlwaysFalsy) | (dict[Unknown | str, Unknown | str | dict[str, str]] & ~AlwaysFalsy) | (Any & ~AlwaysFalsy) | ... omitted 3 union elements`
run_agent.py:2389: [invalid-argument-type] invalid-argument-type: Argument to function `_is_oauth_token` is incorrect: Expected `str`, found `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | ... omitted 4 union elements`
run_agent.py:12328: [invalid-argument-type] invalid-argument-type: Argument to function `_is_oauth_token` is incorrect: Expected `str`, found `str | dict[Unknown | str, Unknown | str | dict[str, str]] | Any | ... omitted 3 union elements`
tests/agent/test_codex_cloudflare_headers.py:163: [unresolved-attribute] unresolved-attribute: Attribute `startswith` is not defined on `dict[str, str]` in union `Unknown | str | dict[str, str]`
run_agent.py:2160: [invalid-argument-type] invalid-argument-type: Argument to function `query_ollama_num_ctx` is incorrect: Expected `str`, found `(str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | (Any & ~AlwaysFalsy) | ... omitted 4 union elements`
tests/run_agent/test_provider_attribution_headers.py:130: [unsupported-operator] unsupported-operator: Operator `not in` is not supported between objects of type `Literal["X-OpenRouter-Cache"]` and `Unknown | str | dict[str, str] | ... omitted 3 union elements`
run_agent.py:6296: [invalid-argument-type] invalid-argument-type: Argument to function `_codex_cloudflare_headers` is incorrect: Expected `str`, found `Unknown | str | dict[str, str] | ... omitted 3 union elements`
tests/agent/test_codex_cloudflare_headers.py:163: [unresolved-attribute] unresolved-attribute: Attribute `get` is not defined on `str & ~AlwaysFalsy`, `int & ~AlwaysFalsy` in union `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | ... omitted 3 union elements`
run_agent.py:2386: [invalid-argument-type] invalid-argument-type: Argument to function `build_anthropic_client` is incorrect: Expected `str`, found `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | ... omitted 4 union elements`
tests/run_agent/test_provider_attribution_headers.py:131: [unsupported-operator] unsupported-operator: Operator `not in` is not supported between objects of type `Literal["X-OpenRouter-Cache-TTL"]` and `Unknown | str | dict[str, str] | ... omitted 3 union elements`

✅ Fixed issues (9):

Rule	Count
`invalid-argument-type`	8
`unresolved-attribute`	1

First entries

run_agent.py:12328: [invalid-argument-type] invalid-argument-type: Argument to function `_is_oauth_token` is incorrect: Expected `str`, found `str | dict[Unknown, Unknown] | Any | ... omitted 4 union elements`
run_agent.py:2389: [invalid-argument-type] invalid-argument-type: Argument to function `_is_oauth_token` is incorrect: Expected `str`, found `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | ... omitted 5 union elements`
run_agent.py:2160: [invalid-argument-type] invalid-argument-type: Argument to function `query_ollama_num_ctx` is incorrect: Expected `str`, found `(str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | (Any & ~AlwaysFalsy) | ... omitted 5 union elements`
run_agent.py:2386: [invalid-argument-type] invalid-argument-type: Argument to function `build_anthropic_client` is incorrect: Expected `str`, found `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | ... omitted 5 union elements`
run_agent.py:6296: [invalid-argument-type] invalid-argument-type: Argument to function `_codex_cloudflare_headers` is incorrect: Expected `str`, found `Unknown | str | dict[str, str] | dict[Unknown, Unknown] | Divergent`
run_agent.py:6467: [invalid-argument-type] invalid-argument-type: Argument to function `build_anthropic_client` is incorrect: Expected `str`, found `str | dict[Unknown, Unknown] | Any | ... omitted 4 union elements`
tests/agent/test_codex_cloudflare_headers.py:163: [unresolved-attribute] unresolved-attribute: Attribute `get` is not defined on `str & ~AlwaysFalsy` in union `(Unknown & ~AlwaysFalsy) | (str & ~AlwaysFalsy) | (dict[str, str] & ~AlwaysFalsy) | dict[Unknown, Unknown] | Divergent`
run_agent.py:12331: [invalid-argument-type] invalid-argument-type: Argument to function `len` is incorrect: Expected `Sized`, found `(str & ~AlwaysFalsy) | (dict[Unknown, Unknown] & ~AlwaysFalsy) | (Any & ~AlwaysFalsy) | ... omitted 4 union elements`
run_agent.py:2437: [invalid-argument-type] invalid-argument-type: Argument to function `get_model_context_length` is incorrect: Expected `str`, found `str | dict[str, str] | Any | ... omitted 4 union elements`

Unchanged: 3934 pre-existing issues carried over.

Diagnostics are surfaced as warnings — this check never fails the build.

@counterposition

…s empty (NousResearch#21245) When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR NousResearch#20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

@counterposition

…s empty (NousResearch#21245) When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR NousResearch#20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

@counterposition

…s empty (NousResearch#21245) When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR NousResearch#20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

@counterposition

…s empty (NousResearch#21245) When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR NousResearch#20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

@counterposition

…s empty (NousResearch#21245) When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR NousResearch#20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

@counterposition

…s empty (NousResearch#21245) When the provider rejects a request (e.g. invalid model slug like '--provider nous --model kimi-k2.6' where the valid slug is 'moonshotai/kimi-k2.6'), run_conversation() returns {failed: True, error: <detail>, final_response: None}. The TUI gateway and one-shot CLI mode both dropped the error on the floor and emitted an empty turn, so the user saw a blank response with no indication that anything went wrong. Mirror the interactive CLI's existing pattern (cli.py:9832): when final_response is empty AND (failed|partial) is set AND error is populated, surface 'Error: <detail>' as the visible text. Leaves the None-with-no-error path and the '(empty)' sentinel path untouched — an empty successful turn still renders empty, and existing sentinel handlers keep owning their lane. Reported by @counterposition in PR NousResearch#20873; taking a minimal fix rather than the broader structured-failure refactor proposed there.

teknium1 merged commit 6e46f99 into main May 7, 2026
8 of 11 checks passed

teknium1 deleted the fix/tui-surface-backend-errors branch May 7, 2026 12:53

teknium1 mentioned this pull request May 7, 2026

fix: display structured agent result failures #20873

Closed

alt-glitch added type/bug Something isn't working P2 Medium — degraded but workaround exists comp/tui Terminal UI (ui-tui/ + tui_gateway/) comp/cli CLI entry point, hermes_cli/, setup wizard labels May 7, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(tui): surface backend error as visible text when final_response is empty#21245

fix(tui): surface backend error as visible text when final_response is empty#21245
teknium1 merged 1 commit into
mainfrom
fix/tui-surface-backend-errors

teknium1 commented May 7, 2026

Uh oh!

Uh oh!

github-actions Bot commented May 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

teknium1 commented May 7, 2026

Summary

Repro

Changes

What stays the same

Validation

Uh oh!

Uh oh!

github-actions Bot commented May 7, 2026

🔎 Lint report: fix/tui-surface-backend-errors vs origin/main

ruff

ty (type checker)

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

🔎 Lint report: `fix/tui-surface-backend-errors` vs `origin/main`