| name | second-opinion |
| description | Second opinion: request consultant advice or set its mode. |
Second Opinion
Use this skill to invoke $consultant or to change Qwen-line consultant mode.
Toggle keys
Manage these keys in .qwen/.agents-mode.yaml:
-
Legacy .qwen/.agents-mode is compatibility input only. Resolve Qwen overlay state in this order: local .qwen/.agents-mode.yaml, local legacy .qwen/.agents-mode, global ~/.qwen/.agents-mode.yaml, then global legacy ~/.qwen/.agents-mode. Normalize whichever file supplied the effective config into the canonical .yaml path in the same scope, do not recreate any legacy file, and do not synthesize a local override on read alone.
-
consultantMode
-
delegationMode
-
parallelMode
-
mcpMode
-
preferExternalWorker
-
preferExternalReviewer
-
externalProvider
-
externalPriorityProfile
-
reserveResolver
-
externalPriorityProfiles
-
externalOpinionCounts
-
externalModelMode
-
externalCodexProfile
Qwen-line rules:
externalProvider: auto resolves through the active named priority profile, not a Qwen-line default provider
externalPriorityProfile defaults to balanced
reserveResolver binds symbolic reserve to claude-sonnet, claude-wrapper, wrapper:<command>, or disabled
- the shipped
balanced profile is production-only and keeps auto routing on codex | claude
- explicit providers are
codex, claude, gemini, and qwen
externalProvider: gemini and externalProvider: qwen are explicit example-only overrides; both are WEAK MODEL / NOT RECOMMENDED
externalModelMode is the shared cross-provider model policy: runtime-default leaves the resolved provider on its runtime default model/profile, while pinned-top-pro starts on the strongest documented provider-native production path for the resolved provider
externalCodexProfile: default inherits externalModelMode when Codex is selected or auto-resolved; gpt-5.6-sol-max requests model gpt-5.6-sol with model_reasoning_effort = "max" for higher-complexity/hard lanes; gpt-5.6-terra selects the balanced Codex model tier (a distinct model, model_reasoning_effort = "high", not an effort downgrade) and must be verified against the installed Codex runtime; gpt-5.6-sol-xhigh (shipped as default in the Codex/Claude packs) pins model gpt-5.6-sol with model_reasoning_effort = "xhigh" regardless of externalModelMode, symmetric to Claude's opus-xhigh
reserve matters only when an advisory/review profile order reaches that symbolic supplemental candidate and is bound through reserveResolver
- if a repository wants Qwen participation for a specific example lane, express that through a scalar explicit provider override, not a profile entry
- same-provider Qwen routing must be explicit; ordinary
auto must still avoid self-bounce
- preserve unknown keys and keep
reserveResolver plus the profile/count keys in expanded multi-key form rather than collapsing them into a consultant-only shape
parallelMode is the general helper fan-out rule across internal and external lanes
externalOpinionCounts is lane-specific; when a lane asks for more than one opinion, the lead may invoke the matching external skill repeatedly and aggregate fail closed on top of parallelMode
Toggle actions
enable -> consultantMode: external
internal -> consultantMode: internal
disable -> consultantMode: disabled
status -> read and normalize .qwen/.agents-mode.yaml, then print the current resolved values
- If local
.qwen/.agents-mode.yaml is missing, read local legacy .qwen/.agents-mode as compatibility input only; if both local files are missing, fall back through pack-local global ~/.qwen/.agents-mode.yaml, pack-local global legacy ~/.qwen/.agents-mode, then the shared cross-pack global ~/.agents-mode.yaml (alongside ~/.claude.json), before applying built-in defaults (each key resolves to the highest layer that defines it; layers compose, they do not replace each other wholesale) and reporting status
Preserve unknown keys on write and normalize comment-free, partial, or older-layout files to the current canonical format on read. Keep one key per line with inline allowed-value comments. Legacy .qwen/.agents-mode is compatibility input only and must not be recreated.
Advisory run
When not in toggle mode:
- read and normalize the current toggle state
- collect the bounded question and accepted context
- invoke
$consultant
- return the memo
Consultant output remains non-blocking.