sed -n '1,220p' /Users/prasanna/.codex/plugins/cache/openai-curated/github/603a6e80/skills/github/SKILL.md

---
name: github
description: Triage and orient GitHub repository, pull request, and issue work through the connected GitHub app. Use when the user asks for general GitHub help, wants PR or issue summaries, or needs repository context before choosing a more specific GitHub workflow.
---

# GitHub

## Overview

Use this skill as the umbrella entrypoint for general GitHub work in this plugin. It sho…

find agent_memory -maxdepth 3 -type f -print

agent_memory/releases/release-v0.1.0-2026-01-22.md
agent_memory/releases/release-v0.3.5-hotfix-2026-04-28.md
agent_memory/releases/release-v0.4.0-2026-05-15.md
agent_memory/releases/release-v0.3.0-2026-03-31.md
agent_memory/releases/release-v0.6.0-2026-06-22.md
agent_memory/releases/release-v0.6.2-prep-2026-06-30.md
agent_memory/releases/release-v0.2.0-2026-02-28.md
agent_memory/releases/release-v…

rg -n \"PR #84|#84|pull request 84|concurrent file processing|60 MB|Wikipedia|v0\\.5\\.0|release notes|benchmark|wall-clock|hardware\" agent_memory .

agent_memory/email/sponsor-outreach-bytebase.md:76:  - v0.5.0 shipped today: concurrent file processing, which makes batch
agent_memory/email/sponsor-outreach-bytebase.md:77:    conversions dramatically faster on multi-core hardware. Hiroshi
agent_memory/email/sponsor-outreach-bytebase.md:79:    profiling and the implementation (PR #84). The benchmark results on
agent_memory/email/sponsor-outreach…

git remote -v && git status --short --branch

origin	git@github.com:supermemoryai/smfs.git (fetch)
origin	git@github.com:supermemoryai/smfs.git (push)
## feat/release-rl-improvements...origin/feat/release-rl-improvements [ahead 1]
 M ../../../../../../crates/smfs-core/src/agent_hint.rs

sed -n '1,260p' agent_memory/pr_threads/pr-084-concurrent-processing.md

# PR #84: perf(cli): concurrent file processing for batched conversions (closes #41)

**Branch:** `hiroshi-n/concurrent-processing` → `main`
**Opened:** 2026-05-30 by @hiroshi-n
**Merged:** 2026-06-08 by @lior-bz
**Diff:** 14 files changed, +423 −158
**Status:** Merged ✅ — v0.5.0

---

## Description

### Problem

Batched conversions on corpora larger than ~50MB showed a severe performance regress…

sed -n '1,180p' agent_memory/notes/retro-june-2026.md

# June 2026 retro — v0.5.0, v0.6.0, Bytebase, v0.7 kickoff

**Date:** 2026-06-30

---

Two releases in one month. I don't think that's happened before — not real releases, anyway. June was heavier than I expected in a good way, and I want to capture it before I fall asleep at my desk.

Related: → [memory/profiles/lior-brennan-zhou.md](../memory/profiles/lior-brennan-zhou.md)

---

## Wins

**v0.5.…

sed -n '1,260p' agent_memory/releases/release-v0.5.0-2026-06-08.md

# kitabi v0.5.0 (2026-06-08) — Concurrent file processing

> **Breaking change:** plugin-API ABI version bumped from 1 → 2. Plugin authors must rebuild against `kitabi-plugin-api` 0.5.x. See [CHANGELOG-PLUGIN-API.md](../../CHANGELOG-PLUGIN-API.md) and the migration section below.

---

## Highlights 🚀

- **5× throughput improvement** on multi-file batches. A 60 MB Wikipedia subset that took 7.4 s…

sed -n '190,235p' agent_memory/code/ARCHITECTURE.md

| 2    | parse error                              |
| 3    | ambiguity (the fail-fast rule fired)     |
| 4    | plugin error (load, ABI, capability)     |
| 5    | I/O error                                |

Diagnostics go through `miette` 7.x. Ambiguity errors point at source spans with carets — when you see one and it's *not* helpful, that is a bug; please file it.

## Configuration & Logging

…

sed -n '260,560p' agent_memory/pr_threads/pr-084-concurrent-processing.md

For reference, here is the full hyperfine sweep I ran across five corpus sizes. All runs on M2 Max, 12 physical cores, 32 GB RAM, macOS Sequoia 15.3. Corpus is the same Wikipedia markdown subset across all sizes (subsampled uniformly by file count). Each benchmark is 10 warm runs after 3 cold discards.

The scaling is essentially linear. At 1MB the overhead of spawning tokio tasks is visible as a…

sed -n '1,260p' agent_memory/issues/issue-097-bench-suite-flaky.md

# Bench suite is flaky on GitHub Actions (high variance, false regressions)

**Issue #97**
**Opened:** 2026-06-10
**Reporter:** @hiroshi-n
**Labels:** `infra`, `bench`, `ci`, `open`
**Status:** open

---

## Body

@hiroshi-n — 2026-06-10 23:14 JST

Filing this so we have a tracking issue. Short version: our bench job on GitHub Actions is too noisy to be useful for catching regressions, and over th…

sed -n '1,260p' agent_memory/notes/retro-may-2026.md

# May 2026 retro — plugin system, ABI break, ban, Kavya

**Date:** 2026-05-31

Predecessor: [retro-april-2026.md](retro-april-2026.md)

May was the most eventful month since the /r/rust post in February. Shipped the biggest release yet, had the worst self-inflicted bug yet, gained a contributor I'm genuinely excited about, watched a star counter I'd convinced myself I didn't care about tick over 1…

rg -n \"3,142|3142|1,200|1200|max-concurrency|--jobs|M2 Max|M1 Pro|280 ms|278\\.3|Wikipedia subset|wikipedia-60mb|60MB|60 MB\" agent_memory

agent_memory/email/conf-invite-rustconf-2026.md:74:- Travel: economy airfare reimbursed up to $1,200 USD; hotel for 4 nights (single room, conference hotel block)
agent_memory/releases/release-v0.5.0-2026-06-08.md:9:- **5× throughput improvement** on multi-file batches. A 60 MB Wikipedia subset that took 7.4 seconds on v0.4.x now processes in **280 ms** on an M2 Max. The wall-clock gains are even …

sed -n '1,120p' agent_memory/issues/issue-041-performance-regression-50mb.md

# Issue #41 — Performance regression on large files (>50MB) — converter slows ~5× vs v0.2

**Repository:** lior-bz/kitabi
**Opened:** 2026-03-22 by @hiroshi-n
**Closed:** 2026-06-08 by @lior-bz (closed via PR #84 merge)
**Milestone:** v0.5.0
**Labels:** `perf` `regression` `tree-sitter`
**Assignee:** @hiroshi-n

Cross-references: ARCHITECTURE.md · adrs/ADR-002-parsing-tree-sitter.md · pr_threads/p…

sed -n '140,270p' agent_memory/issues/issue-041-performance-regression-50mb.md

---

## @nadiap — 2026-03-26 09:02 EEST

Thank you. I'll add a note in CONTRIBUTING.md about checking diagnostic output as part of perf work.

---

## @hiroshi-n — 2026-04-02 03:27 JST

PR #62 merged last week. Bumpalo is in. Re-baselining on master:

| Build              | 60MB single-file | Notes                 |
|--------------------|------------------|-----------------------|
| v0.3.0 (relea…

sed -n '120,155p' agent_memory/releases/release-v0.5.0-2026-06-08.md

Log ordering across concurrent files is buffered per-file. If you are parsing kitabi output programmatically, use `--output-format=json` (stable since v0.3.0) instead of scraping the human log.

---

## Bug fixes

- **#41** — Performance regression on files >50 MB. Hiroshi reported in early April that processing a 60 MB file had gotten ~4× slower between v0.3.0 and v0.4.0, tracing to a quadratic r…

sed -n '1,190p' agent_memory/memory/community/hiroshi-nakamura.md

# Profile — Hiroshi Nakamura (@hiroshi-n)

**Pronouns:** he/him
**Location:** Tokyo, Japan
**GitHub:** @hiroshi-n
**Email (security correspondence):** hiroshi.nakamura@example.jp
**PGP key:** registered with the project for security@ communication

---

## Identity and background

Hiroshi Nakamura is a Tokyo-based engineer with a background in benchmarking, performance engineering, and security re…

nl -ba agent_memory/pr_threads/pr-084-concurrent-processing.md | sed -n '25,70p'

    25	Error handling for partial failures: if one or more files fail in a batch, the remaining files complete, failed files emit miette diagnostics with full context, and the process exits with a non-zero code for the batch. This mirrors RFC-001's fail-fast-on-ambiguity rule applied at the batch boundary rather than within a single file.
    26	
    27	### Numbers
    28	
    29	Hardware: Apple M…

nl -ba agent_memory/pr_threads/pr-084-concurrent-processing.md | sed -n '255,270p'

   255	🚀 28  ❤️ 14  ⚡ 18
   256	
   257	---
   258	
   259	## Hiroshi's benchmark note (added to PR body 2026-06-02)
   260	
   261	For reference, here is the full hyperfine sweep I ran across five corpus sizes. All runs on M2 Max, 12 physical cores, 32 GB RAM, macOS Sequoia 15.3. Corpus is the same Wikipedia markdown subset across all sizes (subsampled uniformly by file count). Each benchmark is…

nl -ba agent_memory/releases/release-v0.5.0-2026-06-08.md | sed -n '1,15p'

     1	# kitabi v0.5.0 (2026-06-08) — Concurrent file processing
     2	
     3	> **Breaking change:** plugin-API ABI version bumped from 1 → 2. Plugin authors must rebuild against `kitabi-plugin-api` 0.5.x. See [CHANGELOG-PLUGIN-API.md](../../CHANGELOG-PLUGIN-API.md) and the migration section below.
     4	
     5	---
     6	
     7	## Highlights 🚀
     8	
     9	- **5× throughput improvement** …

nl -ba agent_memory/issues/issue-041-performance-regression-50mb.md | sed -n '188,230p'

   188	PR #84 is up for review. Concurrent file processing, bounded join set, blocking-pool parser, batched writers. Bench numbers on the directory corpus (600 files, 60MB total, M1 Pro — Lior's box, not mine, since he asked):
   189	
   190	| Configuration                     | Mean wall time |
   191	|-----------------------------------|---------------:|
   192	| serial, pre-bumpalo (v0.3.0)    …

nl -ba agent_memory/releases/release-v0.5.0-2026-06-08.md | sed -n '132,148p'

   132	
   133	<details>
   134	<summary>Full bench numbers by corpus size (M2 Max, macOS 14.5, --max-concurrency=8)</summary>
   135	
   136	| Corpus | Files | Total size | v0.4.2 | v0.5.0 | Speedup |
   137	|---|---|---|---|---|---|
   138	| Tiny (synthetic) | 10 | 500 KB | 48 ms | 12 ms | 4.0× |
   139	| Small (personal wiki) | 120 | 8 MB | 410 ms | 88 ms | 4.7× |
   140	| Medium (project docs)…

pwd && ls

/Users/prasanna/Documents/supermemoryai/supermemoryfs/.scratch/eval/eval/workspace/smfs/dp_006/agent_memory
code
email
issues
memory
notes
pr_threads
profile.md
releases

smfs grep "v0.5.0 release notes 60 MB Wikipedia corpus wall clock"

# supermemory semantic search — 10 results for "v0.5.0 release notes 60 MB Wikipedia corpus wall clock"
# searches by meaning across files in this container. usage:
#   grep "natural language query"          search all files
#   grep "query" path/to/dir/              search within directory
# output: <filepath>:<line_start>-<line_end>:<chunk> (chunk text is verbatim from the file)
/memory/profiles…

smfs grep "PR #84 concurrent file processing 60 MB Wikipedia corpus benchmark hardware"

# supermemory semantic search — 10 results for "PR #84 concurrent file processing 60 MB Wikipedia corpus benchmark hardware"
# searches by meaning across files in this container. usage:
#   grep "natural language query"          search all files
/memory/community/hiroshi-nakamura.md:Hiroshi Nakamura architected concurrent file processing in kitabi (PR #84), reducing processing time for large files…

nl -ba pr_threads/pr-084-concurrent-processing.md | sed -n '1,80p'

     1	# PR #84: perf(cli): concurrent file processing for batched conversions (closes #41)
     2	
     3	**Branch:** `hiroshi-n/concurrent-processing` → `main`
     4	**Opened:** 2026-05-30 by @hiroshi-n
     5	**Merged:** 2026-06-08 by @lior-bz
     6	**Diff:** 14 files changed, +423 −158
     7	**Status:** Merged ✅ — v0.5.0
     8	
     9	---
    10	
    11	## Description
    12	
    13	### Pr…

nl -ba releases/release-v0.5.0-2026-06-08.md | sed -n '110,165p'

   110	Log lines from concurrent file processing are now buffered per-file and written atomically when the file finishes processing. The order of completion is not guaranteed (it depends on file size and content), but you will never see lines from two files interleaved. With `--verbose`, each file's log block is prefixed with the filename:
   111	
   112	```
   113	[docs/guide.md] INFO  parsing 14…

nl -ba pr_threads/pr-084-concurrent-processing.md | sed -n '240,310p'

   240	
   241	### @lior-bz — 2026-06-05T16:02:00Z
   242	
   243	Reviewed the updated diff. The JoinSet + Semaphore pattern is clean. Error semantics are correct and match RFC-001's intent at the batch boundary. Log ordering is solved. Exit-code handling is consistent with the rest of the CLI.
   244	
   245	Approving. Numbers speak for themselves.
   246	
   247	---
   248	
   249	### @lior-bz —…

nl -ba issues/issue-041-performance-regression-50mb.md | sed -n '170,205p'

   170	---
   171	
   172	## @hiroshi-n — 2026-04-02 22:08 JST
   173	
   174	Acknowledged. The serial path will be preserved as-is and `--max-concurrency 1` will go through it exactly. Concurrency is opt-in via flag with a sensible default of `num_cpus::get_physical()`.
   175	
   176	PR is going to take me a few weeks — there's a backpressure question on the writer side that I want to get right …

rg -n "3,142|3142|1,200|1200|max-concurrency=8|max-concurrency=12|macOS 14.5|Sequoia 15.3|278.3|280 ms|280ms"

releases/release-v0.5.0-2026-06-08.md:9:- **5× throughput improvement** on multi-file batches. A 60 MB Wikipedia subset that took 7.4 seconds on v0.4.x now processes in **280 ms** on an M2 Max. The wall-clock gains are even larger on machines with 8+ cores.
releases/release-v0.5.0-2026-06-08.md:134:<summary>Full bench numbers by corpus size (M2 Max, macOS 14.5, --max-concurrency=8)</summary>
relea…

smfs grep "max-concurrency 8 12 M2 Max release notes PR 84 benchmark"

/memory/community/hiroshi-nakamura.md:Hiroshi Nakamura uses an M2 Max MacBook Pro (32 GB RAM) and tools like hyperfine and criterion for his performance benchmarking.

/memory/community/hiroshi-nakamura.md:Hiroshi Nakamura architected concurrent file processing in kitabi (PR #84), reducing processing time for large files significantly.

# supermemory semantic search — 10 results for "max-concurren…

smfs grep "1200 files 3142 files 60MB Wikipedia subset release notes PR #84 mismatch"

/memory/community/hiroshi-nakamura.md:Hiroshi Nakamura architected concurrent file processing in kitabi (PR #84), reducing processing time for large files significantly.

/memory/community/drift-owl-42.md:Lior Brennan-Zhou has invited drift_owl_42 to take on a co-maintainer triage role, though drift_owl_42 has not yet accepted.

/memory/profiles/sven-halgrimsson.md:Sven Halgrímsson maintains a nar…

nl -ba releases/release-v0.5.0-2026-06-08.md | sed -n '1,25p'

     1	# kitabi v0.5.0 (2026-06-08) — Concurrent file processing
     2	
     3	> **Breaking change:** plugin-API ABI version bumped from 1 → 2. Plugin authors must rebuild against `kitabi-plugin-api` 0.5.x. See [CHANGELOG-PLUGIN-API.md](../../CHANGELOG-PLUGIN-API.md) and the migration section below.
     4	
     5	---
     6	
     7	## Highlights 🚀
     8	
     9	- **5× throughput improvement** …

nl -ba issues/issue-041-performance-regression-50mb.md | sed -n '220,265p'

   220	## @hiroshi-n — 2026-06-08 18:34 JST
   221	
   222	One last note for the record, since this thread will get linked from retros and CONTRIBUTING.md:
   223	
   224	The regression was a known cost of the parser strategy; we paid it intentionally. ADR-002 chose tree-sitter for incremental reparse, IDE friendliness, and structural correctness on weird inputs. The ~5× single-file slowdown was a…

nl -ba pr_threads/pr-084-concurrent-processing.md | sed -n '140,160p'

   140	
   141	---
   142	
   143	### @svenh — 2026-06-01T08:47:00Z
   144	
   145	**[inline — `src/cli/batch.rs:203`]**
   146	
   147	Backpressure. Right now, if someone passes `--max-concurrency=512` on a system with 8 physical cores, you'll spawn 512 concurrent tasks immediately on a corpus with enough files. The JoinSet doesn't have a built-in concurrency cap — tasks are only bounded by how f…

nl -ba code/ARCHITECTURE.md | sed -n '200,230p'

   200	- **Logging**: `tracing` + `tracing-subscriber`. Default is `WARN`+. `--quiet` (PR #79, @kavya-i) suppresses `INFO`. `RUST_LOG` is honored and overrides everything.
   201	
   202	## CLI Surface
   203	
   204	Built on clap v4 derive macros. Commands:
   205	
   206	- `kit convert <input> [-o <output>] [--from <fmt>] [--to <fmt>]`
   207	- `kit check <input>` — parse-only, useful in CI
   2…