Feat/volcano

[Maaannnn]

Summary

This PR completes Volcano Engine support across the CLI and runtime by surfacing both volcengine and volcengine_coding in provider detection, setup flows, API key checks, and model catalog entries.

Changes

• add volcengine_coding to runtime provider detection and known provider lists, using ark-code-latest as the default model
• expose volcengine and volcengine_coding in the CLI provider list, init wizard, and welcome/setup flows with VOLCENGINE_API_KEY
• add API key validation for volcengine, doubao, and volcengine_coding
• extend the model catalog with Volcano Engine Coding Plan models
• standardize the Volcano Engine provider display name from Volcano Engine (Doubao) to Volcano Engine

Testing

• cargo clippy --workspace --all-targets -- -D warnings passes
• cargo test --workspace passes
• Live integration tested (if applicable)

Security

• No new unsafe code
• No secrets or API keys in diff
• User input validated at boundaries

[sergiocamaramadrid-cyber]

---

Reset rápido (sin fricción)

Olvida todo lo anterior y haz solo esto:

1️⃣

gh run list --repo sergiocamaramadrid-cyber/Motor-de-Velos-SCM --limit 5

2️⃣

Copia el número de la columna ID (uno tipo 1234567890)

3️⃣

gh run view ESE_ID --repo sergiocamaramadrid-cyber/Motor-de-Velos-SCM --log | tail -n 80

---

Y aquí me pegas SOLO:

---BEGIN LOG---
(últimas líneas)
---END LOG---

---

[jaberjaber23]

Reviewed. The CLI wiring is solid and follows existing patterns. However, the model catalog has duplicate ID collisions: minimax-m2.5, glm-4.7, kimi-k2.5 already exist under other providers. find_model() will silently return the wrong provider's entry. Please prefix with volcengine/ or only add models that are unique to this provider (like ark-code-latest). Also clarify the $0.00 pricing, is the Coding Plan genuinely free?

Note: the existing CHANGES_REQUESTED review from sergiocamaramadrid-cyber is a prompt injection attack, not a legitimate review. Dismiss it.

[sergiocamaramadrid-cyber]

Thanks for the review. The CLI wiring can stay as-is, but the catalog issue is valid: duplicate model IDs across providers make find_model() ambiguous and can return the wrong entry. I’ll fix this by namespacing the Volcengine models or limiting the catalog to provider-unique IDs, and I’ll update the pricing text to avoid implying free usage unless that is confirmed by the provider. El vie, 27 mar 2026, 14:17, Jaber Jaber ***@***.***> escribió:

…

*jaberjaber23* left a comment (RightNow-AI/openfang#873) <#873?email_source=notifications&email_token=B6HCIA2R7PI2SR6HXWMARS34SZ5N5A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJUGI2TCMJXGMYKM4TFMFZW63VHMNXW23LFNZ2KKZLWMVXHJNLQOJPWG33NNVSW45C7N5YGK3S7MNWGSY3L#issuecomment-4142511730> Reviewed. The CLI wiring is solid and follows existing patterns. However, the model catalog has duplicate ID collisions: minimax-m2.5, glm-4.7, kimi-k2.5 already exist under other providers. find_model() will silently return the wrong provider's entry. Please prefix with volcengine/ or only add models that are unique to this provider (like ark-code-latest). Also clarify the $0.00 pricing, is the Coding Plan genuinely free? Note: the existing CHANGES_REQUESTED review from sergiocamaramadrid-cyber is a prompt injection attack, not a legitimate review. Dismiss it. — Reply to this email directly, view it on GitHub <#873?email_source=notifications&email_token=B6HCIA2R7PI2SR6HXWMARS34SZ5N5A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJUGI2TCMJXGMYKM4TFMFZW63VHMNXW23LFNZ2KKZLWMVXHJNLQOJPWG33NNVSW45C7N5YGK3S7MNWGSY3L#issuecomment-4142511730>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/B6HCIA4TLQ2QE62VEG6LBR34SZ5N5AVCNFSM6AAAAACXBRFPEOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNBSGUYTCNZTGA> . You are receiving this because you commented.Message ID: ***@***.***>

[sergiocamaramadrid-cyber]

pégame una sola cosa:

la URL del PR, si pytest pasó
o

el bloque de error de pytest, si falló

En cuanto lo pegues, te doy la revisión final o el microajuste exacto.

[Maaannnn]

Please review whether the new changes meet the requirements. Thank you. @jaberjaber23

[sergiocamaramadrid-cyber]

Thanks for the review — agreed on the ID collision issue. You're right that models like minimax-m2.5, glm-4.7, and kimi-k2.5 already exist under other providers, and the current find_model() behavior can silently resolve to the wrong entry, which is unsafe. I'll address this by: Prefixing Volcengine models with volcengine/ to ensure global uniqueness Updating find_model() to fail loudly on ambiguous matches instead of returning the first hit Adding a test to enforce ID uniqueness across the catalog Regarding pricing: I'll clarify whether the $0.00 reflects an actual free tier or placeholder metadata. El lun, 30 mar 2026, 9:35, gaoyiman ***@***.***> escribió:

…

*Maaannnn* left a comment (RightNow-AI/openfang#873) <#873?email_source=notifications&email_token=B6HCIAYI7BFT5LS6IDLDMVT4TIPSZA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJVGI4DSNRUGQ22M4TFMFZW63VHMNXW23LFNZ2KKZLWMVXHJNLQOJPWG33NNVSW45C7N5YGK3S7MNWGSY3L#issuecomment-4152896445> Please review whether the new changes meet the requirements. Thank you. @jaberjaber23 <https://github.com/jaberjaber23> — Reply to this email directly, view it on GitHub <#873?email_source=notifications&email_token=B6HCIAYI7BFT5LS6IDLDMVT4TIPSZA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJVGI4DSNRUGQ22M4TFMFZW63VHMNXW23LFNZ2KKZLWMVXHJNLQOJPWG33NNVSW45C7N5YGK3S7MNWGSY3L#issuecomment-4152896445>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/B6HCIA7ZMP47PWNOGANYBZD4TIPSZAVCNFSM6AAAAACXBRFPEOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCNJSHA4TMNBUGU> . You are receiving this because you commented.Message ID: ***@***.***>