feat(composio): promote GitHub from catalog-only to native memory provider by justinhsu1477 · Pull Request #2413 · tinyhumansai/openhuman

justinhsu1477 · 2026-05-21T05:31:34Z

Summary

Adds GitHubProvider next to the existing gmail / notion / slack / clickup providers — GitHub becomes the fifth native memory-ingest provider in composio/providers/. Until now github/mod.rs self-declared "curated tool catalog only — no native ComposioProvider implementation yet"; this PR closes that gap.
Implementation mirrors the ClickUp (feat(composio): add ClickUp provider for Memory Tree ingest #2291, merged) and Linear (feat(composio): add Linear provider for Memory Tree ingest #2402, in review) providers 1:1 — same SyncState semantics, same persist_single_item ingest path, same daily-budget discipline, same "fetch-what-the-user-sees" assignee-scoped fetch.
Privacy posture: only issues where the connected user is the assignee are pulled, never the whole watched-repos issue graph. The assignee:<viewer_login> qualifier is constructed inside the provider — never accepted from a caller — so the boundary can't be tunnelled around.

Problem

composio/providers/ already has working memory-ingest providers for gmail, notion, slack, clickup, and (pending merge of #2402) linear. github/ was structurally one level lower — only tools.rs populated, no ComposioProvider impl. For an open-source AI assistant whose own contributors live on GitHub, that's the highest-leverage missing provider: every maintainer / reviewer / external contributor can dogfood it within minutes of merge.

Solution

Promote composio/providers/github/ to a full provider by adding the three companion files that the other native providers all have:

src/openhuman/composio/providers/github/
  mod.rs        UPDATED  — declares provider/sync/tests modules,
                           drops "no native impl yet" disclaimer
  tools.rs      KEEP     — GITHUB_CURATED unchanged (already includes
                           GITHUB_GET_AUTHENTICATED_USER and
                           GITHUB_SEARCH_ISSUES which is all the sync
                           path needs; no new actions added)
  provider.rs   NEW      — impl ComposioProvider for GitHubProvider
  sync.rs       NEW      — payload helpers (extract_issues,
                           extract_issue_title, extract_issue_updated,
                           extract_repo_qualified_identifier,
                           extract_viewer_login, extract_viewer_id)
  tests.rs      NEW      — 18 unit tests

Plus five registration touchpoints (same shape ClickUp landed in #2291 and Linear in #2402):

composio/providers/mod.rs::has_native_provider: add "github".
composio/providers/mod.rs::native_provider_sync_interval: add "github" => ….
composio/providers/mod.rs::catalog_for_toolkit: no change — already points to github::GITHUB_CURATED.
composio/providers/registry.rs::init_default_providers: add register_provider(Arc::new(GitHubProvider::new())).
composio/providers/descriptions.rs: update GitHub description to mention Memory Tree sync.

Sync model (mirroring ClickUp / Linear)

SyncState::load("github", connection_id) from the shared KV store.
Daily request budget check (DEFAULT_DAILY_REQUEST_LIMIT = 500).
Resolve viewer login via GITHUB_GET_AUTHENTICATED_USER — needed for the assignee:<login> search filter.
Re-check budget between viewer probe and search call (CodeRabbit lesson from feat(composio): add ClickUp provider for Memory Tree ingest #2291 / feat(composio): add Linear provider for Memory Tree ingest #2402 — never burn a list call when the probe just spent the last budget slot).
Page through GITHUB_SEARCH_ISSUES with query is:issue assignee:<viewer_login> sort:updated-desc. GitHub's Search API caps results at 1000 so pagination is naturally bounded; we also stop early on cursor boundary or short page (< per_page).
Per issue, persist as one memory document via persist_single_item. Dedupe by composite issue_id@updated_at so edits re-ingest.
Advance cursor to newest updated_at, record last_sync_at_ms, save state.

Transport-error path in the pagination loop persists state before propagating the error (CodeRabbit's other lesson from #2402: a flap mid-pagination must not roll back budget accounting).

Source-id convention

composio-github-issue-<global_id> — GitHub's id is globally unique across all of GitHub so it's a stable upsert key. Document title surfaces the canonical owner/repo#number form (e.g. GitHub tinyhumansai/openhuman#2408: …) so search hits read the way contributors refer to issues in conversation.

Curated tool catalog

GITHUB_CURATED (already in github/tools.rs from the catalog-only era) is unchanged. GITHUB_GET_AUTHENTICATED_USER and GITHUB_SEARCH_ISSUES were already curated — the minimum the sync path needs — so this PR adds zero new tool actions.

Submission Checklist

Tests added or updated — 18 new unit tests cover sync helpers (results across SEARCH/LIST envelope shapes, title / updated extraction, owner/repo#number composition from both repository.full_name and repository_url shapes, viewer-login strict extractor that refuses to fall back to non-login fields, viewer-id metadata extraction), trait metadata stability, capability matrix registration (capability_matrix_includes_github_as_native_memory_provider), and default_impl_matches_new (observable equivalence — not a no-op test).
Diff coverage ≥ 80% — new code is overwhelmingly the sync() async happy path (covered behind a Composio ProviderContext the existing test harness doesn't stand up — same as gmail / notion / slack / clickup / linear tests don't exercise the live sync() end-to-end either). Helper layer is unit-tested directly.
N/A: Coverage matrix updated — extends the existing "Composio memory provider" capability row; no new matrix feature row.
N/A: All affected feature IDs from the matrix are listed — extending an existing capability.
No new external network dependencies introduced — all GitHub API access goes through the existing Composio backend / direct client.
N/A: Manual smoke checklist updated — no release-cut surface changes; new ingest path is feature-flagged behind "user has a GitHub Composio connection".
Linked issue closed via Closes #2408 in ## Related.

Impact

Runtime/platform impact: desktop core only (Rust). No Tauri shell, no frontend changes.
Compatibility impact: strictly additive. Existing providers, their SyncState KV namespaces (composio-sync-state keyed by (toolkit, connection_id)), their registered tool catalogs, and all catalog-only toolkits are unchanged. The github tool catalog is unchanged — only adds the missing provider impl around it.
Performance impact: bounded — MAX_PAGES_PER_SYNC = 20, PAGE_SIZE = 50 steady-state (100 for the initial backfill), DailyBudget = 500 req/day, and GitHub's own 1000-result Search cap. Per sync pass: 1 viewer probe + up to 10 search pages.
Security impact: assignee-scoped fetch (assignee:<viewer_login>) prevents accidental ingest of other contributors' issues. Viewer-login extractor is strict (no fallback to id / name fields) so the filter can't silently scope to the wrong identity. Composio handles credentials; no new secret-handling code.

Closes Add GitHub as a Composio memory provider (joining gmail / notion / slack / clickup / linear) #2408
Closest templates: composio/providers/clickup/ (feat(composio): add ClickUp provider for Memory Tree ingest #2291, merged), composio/providers/linear/ (feat(composio): add Linear provider for Memory Tree ingest #2402, in review).
Provider trait: composio/providers/traits.rs.
Shared sync state: composio/providers/sync_state.rs.

AI Authored PR Metadata

Linear Issue

Key: N/A
URL: N/A

Commit & Branch

Branch: feat/github-memory-provider
Commit SHA: (latest on the branch)
Base: upstream/main at fetch time

Validation Run

N/A: pnpm --filter openhuman-app format:check — Rust-only change.
N/A: pnpm typecheck — Rust-only change.
Focused tests: cargo test --lib composio::providers::github (40/40 pass — combines 18 in tests.rs and 22 in sync.rs inline tests); cargo test --lib composio::providers (305/305 pass — no regression on gmail / notion / slack / clickup or any catalog-only toolkit).
Rust fmt/check: cargo fmt --check clean; cargo check --lib clean (pre-existing warnings only); cargo clippy --lib --no-deps no new warnings in composio/providers/github/.
N/A: Tauri fmt/check — no app/src-tauri/src/** changes.

Validation Blocked

N/A

Behavior Changes

Intended behavior change: users with a Composio-connected GitHub account now have their assigned issues periodically ingested into the Memory Tree on the existing 30-minute scheduler cadence, with initial backfill triggered by the ConnectionCreated hook.
User-visible effect: GitHub issue content (title, body, state, labels, assignees, repo, created_at, updated_at as JSON) becomes available to the agent and retrieval layer the same way Gmail / Notion / Slack / ClickUp content already is.

Parity Contract

Legacy behavior preserved: existing gmail / notion / slack / clickup providers are completely untouched. Their SyncState KV namespaces are unchanged. GITHUB_CURATED (already in github/tools.rs) is byte-identical — no new tool actions added or removed.
Guard/fallback/dispatch parity checks: provider follows the existing ComposioProvider trait contract — daily budget, dedup-by-id, cursor-based pagination, idempotent persist_single_item upserts.
catalog_for_toolkit("github") continues to return github::GITHUB_CURATED (was already wired that way pre-PR; only the trait-impl side of github/ is new).

Duplicate / Superseded PR Handling

Duplicate PR(s): none — verified before opening Add GitHub as a Composio memory provider (joining gmail / notion / slack / clickup / linear) #2408. Searched all open / closed PRs and issues for "github provider" / "github memory" before scoping. Zero prior work on a GitHubProvider impl.
Canonical PR: this PR.
Resolution: N/A.

Summary by CodeRabbit

Localization Updates
- Expanded German language translations with new and updated strings for mascot settings, developer menu features, appearance preferences, and onboarding/workspace functionality.

…vider Adds `GitHubProvider` next to the existing `gmail` / `notion` / `slack` / `clickup` providers, joining them as the fifth native memory-ingest provider in `composio/providers/`. Until now `github/mod.rs` declared itself "curated tool catalog only — no native ComposioProvider implementation yet"; this change closes that gap so the connected user's assigned GitHub issues stream into the Memory Tree on the periodic scheduler. Implementation mirrors the ClickUp (tinyhumansai#2291, merged) and Linear (tinyhumansai#2402, in review) providers 1:1 — same `SyncState` semantics, same `persist_single_item` ingest path, same daily-budget discipline, same "fetch-what-the-user-sees" assignee-scoped fetch. ## Sync model 1. SyncState load + daily budget gate. 2. Resolve viewer login via GITHUB_GET_AUTHENTICATED_USER. 3. Re-check budget after the probe (per CodeRabbit lesson on tinyhumansai#2291 / tinyhumansai#2402 — never burn a list call when the probe just spent the last budget slot). 4. Page through GITHUB_SEARCH_ISSUES with `is:issue assignee:<login> sort:updated-desc`. GitHub's search caps results at 1000 so pagination is naturally bounded; we also stop early on cursor boundary or short page. 5. Per issue, persist as one memory document via persist_single_item. Composite `issue_id@updated_at` dedup key re-syncs edited issues. 6. Advance cursor to newest updated_at, record last_sync_at_ms, save state. Transport-error path in the pagination loop persists state before returning the error (CodeRabbit lesson on tinyhumansai#2402: a flap mid-pagination must not roll back budget accounting). ## Privacy posture `assignee:<viewer_login>` is constructed inside the provider — never accepted from a caller — so the privacy boundary can't be tunnelled around. Matches the discipline gmail / notion / clickup / linear already follow. ## Source-id convention `composio-github-issue-<global_id>`. GitHub's `id` is globally unique across all of GitHub so it's a stable upsert key. Document title surfaces the canonical `owner/repo#number` form (e.g. `GitHub tinyhumansai#2408: …`) so search hits read the way contributors refer to issues in conversation. ## Curated tool catalog `GITHUB_CURATED` (already in `github/tools.rs` since the catalog-only era) is unchanged — `GITHUB_GET_AUTHENTICATED_USER` and `GITHUB_SEARCH_ISSUES` were already curated, which is the minimum the sync path needs. No new actions added. ## Files Added: - composio/providers/github/provider.rs — GitHubProvider impl (~415) - composio/providers/github/sync.rs — payload helpers (~360) - composio/providers/github/tests.rs — 18 unit tests (~165) Modified: - composio/providers/github/mod.rs — declare new modules + re-export provider; remove stale "no native impl yet" disclaimer - composio/providers/mod.rs — has_native_provider arm + native_provider_sync_interval arm + new regression test `capability_matrix_includes_github_as_native_memory_provider` (catalog_for_toolkit already pointed at github::GITHUB_CURATED; no change needed there) - composio/providers/registry.rs — register_provider in init_default_providers - composio/providers/descriptions.rs — GitHub description updated to mention Memory Tree sync ## Verification - cargo check --lib clean (pre-existing warnings only) - cargo test --lib composio::providers::github 40/40 pass - cargo test --lib composio::providers 305/305 pass (no regression on gmail / notion / slack / clickup) - cargo test --lib capability_matrix_includes_github 1/1 pass - cargo fmt --check clean - cargo clippy --lib --no-deps no new warnings in github/ Closes tinyhumansai#2408

coderabbitai · 2026-05-21T05:31:49Z

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 57fca3bb-a28f-4794-bc10-112c28d10c7d

📥 Commits

Reviewing files that changed from the base of the PR and between b89790c and a164ff9.

📒 Files selected for processing (1)

app/src/lib/i18n/chunks/de-5.ts

💤 Files with no reviewable changes (1)

app/src/lib/i18n/chunks/de-5.ts

📝 Walkthrough

Walkthrough

German locale translations in de-5.ts are updated to remove deprecated custom GIF mascot strings and replace them with new mascot library availability, developer menu, and extended UI/onboarding translation keys.

Changes

German Locale Translations

Layer / File(s)	Summary
Mascot and developer menu translation updates `app/src/lib/i18n/chunks/de-5.ts`	Removes deprecated `settings.mascot.customGifError`, `customGifHeading`, and `customGifLabel` entries; adds `settings.mascot.libraryUnavailable`, `settings.mascot.title`, `settings.developerMenu.composio.*`, and additional UI/onboarding/progress/workspace/calls/migration/settings translation strings in the German translation map.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly related PRs

tinyhumansai/openhuman#2597: Directly modifies the same de-5.ts file, removing and replacing settings.mascot.customGif* keys with new mascot library and developer menu translations.
tinyhumansai/openhuman#2522: Updates mascot "custom GIF" translation entries in de-5.ts in the same settings.mascot.* section.
tinyhumansai/openhuman#2538: Applies parallel mascot "custom GIF" translation updates in the Chinese locale (zh-CN-5.ts).

Suggested labels

feature, rust-core

Suggested reviewers

graycyrus

Poem

🐰 Hallo, liebste Übersetzungen so fein!
Alte Masken-GIFs, weg muss das sein,
Neue Bibliotheken und Menüs erblühen,
Deutsch schreitet voran mit frischen Gefühlen! ✨

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Title check	⚠️ Warning	The PR title references promoting GitHub to a native memory provider, but the file changed is a German i18n translation chunk with no visible connection to composio provider implementation.	The title describes composio/GitHub provider changes, but the only file summary shows i18n translation updates (de-5.ts). Verify the PR includes the actual composio provider code changes or clarify if this is a partial view of the changeset.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

coderabbitai

🧹 Nitpick comments (1)

src/openhuman/composio/providers/github/mod.rs (1)
24-25: ⚡ Quick win

Prefer *_test.rs wiring for extracted module tests.

Since tests were extracted, wire them via a sibling *_test.rs file to match the repository convention.
♻️ Suggested update
 #[cfg(test)]
-mod tests;
+#[path = "github_test.rs"]
+mod tests;
Also rename src/openhuman/composio/providers/github/tests.rs to src/openhuman/composio/providers/github/github_test.rs.

As per coding guidelines: “When extracting Rust tests out of an implementation file, prefer a sibling *_test.rs file wired in with #[cfg(test)] #[path = "..._test.rs"] mod tests;.”
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/openhuman/composio/providers/github/mod.rs` around lines 24 - 25, Replace
the current inline test module wiring (the existing #[cfg(test)] mod tests;) so
it points to a sibling test file using a path attribute and rename the extracted
tests file accordingly: create/rename
src/openhuman/composio/providers/github/tests.rs to github_test.rs and change
the module declaration to use #[cfg(test)] with #[path = "github_test.rs"] mod
tests; so the tests are loaded from the sibling github_test.rs file instead of
inline.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@src/openhuman/composio/providers/github/mod.rs`:
- Around line 24-25: Replace the current inline test module wiring (the existing
#[cfg(test)] mod tests;) so it points to a sibling test file using a path
attribute and rename the extracted tests file accordingly: create/rename
src/openhuman/composio/providers/github/tests.rs to github_test.rs and change
the module declaration to use #[cfg(test)] with #[path = "github_test.rs"] mod
tests; so the tests are loaded from the sibling github_test.rs file instead of
inline.

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 4551cde9-34d5-4fa2-b112-369040495c1a

📥 Commits

Reviewing files that changed from the base of the PR and between 6281aea and b89790c.

📒 Files selected for processing (7)

src/openhuman/composio/providers/descriptions.rs
src/openhuman/composio/providers/github/mod.rs
src/openhuman/composio/providers/github/provider.rs
src/openhuman/composio/providers/github/sync.rs
src/openhuman/composio/providers/github/tests.rs
src/openhuman/composio/providers/mod.rs
src/openhuman/composio/providers/registry.rs

justinhsu1477 · 2026-05-21T19:45:42Z

Status note for anyone scanning this surface:

RFC: Add GitHub as a Composio memory provider (joining gmail / notion / slack / clickup / linear) #2408 (GitHub as Composio memory provider) — mirrors the gmail / notion / slack / clickup / linear provider pattern.
Follow-up: Expand GitHub memory sync from assignee-only to GitHub's involves: qualifier (covers author + assignee + mentions + commenter) #2418 (assignee: → involves: query expansion) layers on top once this lands. @obchain has signalled pickup there.

If anyone has a parallel GitHub-provider implementation in progress or planned, please flag in #2408 first so we can coordinate scope split.

# Conflicts: # src/openhuman/composio/providers/github/mod.rs # src/openhuman/composio/providers/mod.rs # src/openhuman/memory_sync/composio/providers/github/provider.rs # src/openhuman/memory_sync/composio/providers/github/sync.rs # src/openhuman/memory_sync/composio/providers/github/tests.rs

Merge introduced duplicate settings.mascot.customGif* entries.

justinhsu1477 requested a review from a team May 21, 2026 05:31

coderabbitai Bot added the working A PR that is being worked on by the team. label May 21, 2026

coderabbitai Bot reviewed May 21, 2026

View reviewed changes

coderabbitai Bot previously approved these changes May 21, 2026

View reviewed changes

justinhsu1477 mentioned this pull request May 21, 2026

Expand GitHub memory sync from assignee-only to GitHub's involves: qualifier (covers author + assignee + mentions + commenter) #2418

Open

5 tasks

senamakel self-assigned this May 25, 2026

senamakel dismissed coderabbitai[bot]’s stale review via b3117ae May 25, 2026 04:58

fix(i18n): remove duplicate keys in de-5.ts causing TS1117

a164ff9

Merge introduced duplicate settings.mascot.customGif* entries.

coderabbitai Bot added feature Net-new user-facing capability or product behavior. rust-core Core Rust runtime in src/: CLI, core_server, shared infrastructure. labels May 25, 2026

coderabbitai Bot approved these changes May 25, 2026

View reviewed changes

Merge remote-tracking branch 'upstream/main' into pr/2413

8e74681

senamakel merged commit bf2400f into tinyhumansai:main May 25, 2026
28 of 30 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(composio): promote GitHub from catalog-only to native memory provider#2413

feat(composio): promote GitHub from catalog-only to native memory provider#2413
senamakel merged 4 commits into
tinyhumansai:mainfrom
justinhsu1477:feat/github-memory-provider

justinhsu1477 commented May 21, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented May 21, 2026 •

edited

Loading

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Suggested labels

Suggested reviewers

Poem

❌ Failed checks (1 warning)

Uh oh!

coderabbitai Bot left a comment

Uh oh!

justinhsu1477 commented May 21, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

justinhsu1477 commented May 21, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Problem

Solution

Sync model (mirroring ClickUp / Linear)

Source-id convention

Curated tool catalog

Submission Checklist

Impact

Related

AI Authored PR Metadata

Linear Issue

Commit & Branch

Validation Run

Validation Blocked

Behavior Changes

Parity Contract

Duplicate / Superseded PR Handling

Summary by CodeRabbit

Uh oh!

coderabbitai Bot commented May 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Estimated code review effort

Possibly related PRs

Suggested labels

Suggested reviewers

Poem

❌ Failed checks (1 warning)

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

justinhsu1477 commented May 21, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

justinhsu1477 commented May 21, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented May 21, 2026 •

edited

Loading