Add zep-ms-agent-framework integration (Microsoft Agent Framework)#522
Merged
Conversation
27ca1e2 to
9fa7cf9
Compare
5afe848 to
d6bea42
Compare
danielchalef
added a commit
that referenced
this pull request
Jun 18, 2026
The Claude Code Review workflow (added in #530) granted only `pull-requests: read` / `issues: read`. The anthropics/claude-code-action needs Pull requests + Issues write access to post its findings as inline PR comments — with read-only, the run succeeds but posts nothing (observed: `permission_denials_count: 20`, `No buffered inline comments`, zero comments on PRs #522–#527/#529). Per the official docs, the action's GitHub App requires Contents/Issues/Pull requests Read & write, and every posting example uses `pull-requests: write` + `issues: write`. Bump pull-requests and issues to write (contents stays read — review-only, no commits/PRs created). Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
danielchalef
added a commit
that referenced
this pull request
Jun 18, 2026
The code-review plugin only posts findings when invoked with `--comment`; without it the action reviews the PR (spending tokens) but prints to the hidden terminal and posts nothing. Confirmed across PRs #522–#527/#529: runs complete `success` with write perms (post-#531) but log "No buffered inline comments" and leave zero comments. Per the plugin spec ("If --comment was NOT provided, stop here. Do not post any GitHub comments."). Append --comment to the prompt; #531's pull-requests/issues write perms are required for it to post and stay in place. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Memory integration for the Microsoft Agent Framework (Python) via a ContextProvider that persists turns to Zep and injects a Context Block. - ZepContextProvider(ContextProvider): before_run persists the latest user message (thread.add_messages(return_context=True)) and injects the returned Context Block via context.extend_instructions; after_run persists the assistant response. Lazy user/thread creation; Zep failures degrade to memoryless rather than crashing the agent. - Package: integrations/ms-agent-framework/python; dist zep-ms-agent-framework; import zep_ms_agent_framework. Depends on agent-framework-core>=1.8.1 (stable) and zep-cloud>=3.23.0. Python >=3.11. - README, SETUP (Zep signup), runnable example (OpenAIChatClient), 43 mock tests + a live integration test. CI paths-filter entry added. Verified: ruff + ruff format + mypy + pytest (43 passed, 1 skipped) on Python 3.11-3.13; live end-to-end smoke test against Zep + OpenAI passed. See integrations/SPIKE_FINDINGS.md. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…ms-agent-framework - Add shared _text.truncate_message_content helper to cap thread messages at Zep's 4,096-char limit (truncate to ~4,000, never silently drop), warning with lengths only (no content/PII). Apply to both before_run (user message) and after_run (assistant message) so an oversize turn is persisted instead of triggering a swallowed 400. - Track a per-run _user_turn_persisted flag (reset each before_run, not the per-session state dict) and gate after_run on it, so a failed user-turn persist no longer leaves an orphaned assistant-only record. - Guard ONLY `import agent_framework` in __init__; import context_provider outside the try so a broken zep_cloud surfaces its own error rather than being mislabeled as a missing Agent Framework dependency. - Add regression tests for oversize truncation (both hooks), warning content safety, and after_run skipping when the user turn failed. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
97fddc0 to
2fbd9c5
Compare
danielchalef
added a commit
that referenced
this pull request
Jun 18, 2026
The public Claude Code Review step had no timeout, so a hung run sat in_progress at the job default (360 min) instead of failing fast. A review that stalls is an infra hang, not a slow review (observed on PR #522: a run sat unchanged for 48 min until manually cancelled). Cap the step at 10 minutes so a stalled review resolves to a failed check that can be re-run, matching zep-proprietary's claude-review. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This comment has been minimized.
This comment has been minimized.
The python paths-filter used key `agent-framework` → `integrations/agent-framework/python/**`, but the package lives at `integrations/ms-agent-framework/python/` (dist `zep-ms-agent-framework`, import `zep_ms_agent_framework`). The filter never matched, so lint / type-check / tests never ran for this package in CI. The matched filter key flows through as the test-python matrix `package`, and the test-python action resolves `integrations/<package>/python` + import `zep_<package|tr - _>`. So the key must equal the directory name. Correct both key and path to `ms-agent-framework`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
# Conflicts: # .github/workflows/test-integrations.yml
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.
Adds the
zep-ms-agent-frameworkpackage (integrations/ms-agent-framework/python) — Zep memory for the Microsoft Agent Framework.Hook (verified against installed
agent-framework-core1.8.1)ZepContextProvider(ContextProvider)—before_runpersists the latest user message viathread.add_messages(return_context=True)and injects the returned Context Block withcontext.extend_instructions;after_runpersists the assistant response. Attach viacontext_providers=[ZepContextProvider(...)]on theAgent. Lazy user/thread creation; Zep failures never crash the agent.Names / deps
dist
zep-ms-agent-framework, importzep_ms_agent_framework.agent-framework-core>=1.8.1(stable; no--prerelease),zep-cloud>=3.23.0. Python ≥3.11.Ships
README, SETUP.md (Zep signup at getzep.com), runnable example (OpenAIChatClient), 43 mock tests + a live integration test, Makefile, CHANGELOG.
Validation
ruff + ruff format + mypy + pytest (43 passed, 1 skipped) on Python 3.11–3.13; live end-to-end smoke test against Zep + OpenAI passed. Approach in
integrations/SPIKE_FINDINGS.md.🤖 Generated with Claude Code