feat(skills): add int-gemini skill (Google Gemini API offload)#112
Open
mt-alarcon wants to merge 1 commit into
Open
feat(skills): add int-gemini skill (Google Gemini API offload)#112mt-alarcon wants to merge 1 commit into
mt-alarcon wants to merge 1 commit into
Conversation
New integration skill to offload cheap/high-volume tasks from the main
models to the Google Gemini API (AI Studio), via GEMINI_API_KEY.
- Commands: ask (with --allow-sensitive, --fallback, --system-file,
--grounding), scan, route, models, smoke, --version.
- Use cases: classification, summarization, first drafts, and web research
with Google Search grounding — anywhere a large/cheap model fits.
- Confidentiality guard: a PARTIAL guard-rail (documented honestly as such)
that blocks structured PII by default (CPF/CNPJ/card/CEP/email/BRL/phone/
address/PIX/contract/secret) and best-effort name+context. Free-text names
without a role term are a known false-negative, documented + regression-
tested; --allow-sensitive is the conscious override.
- Robust HTTP: retry on 429/5xx, --fallback (flash-lite 503 -> flash) that
never masks a real error, type-safe responses.
- 105 tests, network fully blocked in the suite. smoke is exit-0 + JSON
{overall, steps[], duration_ms}. Stdlib-only; instance/auth via env.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Sorry @mt-alarcon, you have reached your weekly rate limit of 500000 diff characters.
Please try again later or upgrade to continue using Sourcery
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
New integration skill
int-geminito offload cheap / high-volume tasks from the main models to the Google Gemini API (AI Studio), viaGEMINI_API_KEY.Why
A large, low-cost model with a 1M-token window and native Google Search grounding is a natural fit for high-volume, low-stakes work (classification, summarization, first drafts, web research) — letting the primary model focus on reasoning. This skill wraps the Gemini API behind the repo's standard skill conventions.
Design
ask(--allow-sensitive,--fallback,--system-file,--grounding),scan,route,models,smoke,--version.--allow-sensitiveis the conscious override. The SKILL recommends redaction / a no-train tier for inputs that carry user data by nature.--fallback(flash-lite 503 → flash) that reports which model actually answered and never masks a real error; type-safe responses.smoke: exit-0 + JSON{overall, steps[], duration_ms}, including a guard step exercised offline.Tests
105 tests, network fully blocked in the suite (autouse). Includes a documented false-negative corpus (asserts what the guard does not catch, with rationale) so the guard-rail's real limits are visible, not hidden.
Notes
GEMINI_API_KEY) — nothing hardcoded.🤖 Generated with Claude Code