pi-beyin 0.1.11 → 0.1.13

package/CHANGELOG.md

@@ -0,0 +1,73 @@
+# Changelog
+
+## 0.1.13
+
+- No project-specific content in the package: the `/jev-test` self-test payload and the live-test mock
+  are neutral examples, temp fixtures use neutral names, and the vault query comes from
+  `BEYIN_TEST_QUERY` (checks are skipped when it is unset).
+
+## 0.1.12
+
+- README rewritten for the current layout (quick start, verification, scope, privacy, release).
+- Add `CHANGELOG.md` and `SECURITY.md`.
+
+## 0.1.11
+
+- `BEYIN_LANG=en` now also covers the injected context header, the `beyin_jev_*` tool schema and the
+  Jev scoring instruction/criteria. Default stays Turkish.
+- `test/test-skills.ts`: skill inventory, `name:` matching the directory, description budget, no
+  personal paths, `_shared/` references, `pi.skills` manifest. Wired into CI and `run.sh`.
+- `run.sh` warns when the vault's `.agents/skills` duplicates a package skill.
+- Test labels and the live-test narration are English; assertions on product text are unchanged.
+
+## 0.1.10
+
+- English comments and dev output in the CI workflows and the test scripts.
+- `e2e.sh` assertions accept both language builds.
+
+## 0.1.9
+
+- `BEYIN_LANG` (`tr` default, `en`) for the receipt reminder, the distillation offer and the two command
+  descriptions.
+- README: who this is for, the CommandCode provider recommendation, and what the harness patch does on an
+  unfamiliar vault.
+
+## 0.1.8
+
+- Remove version branding from the README, code comments and skills.
+- `THIRD-PARTY-NOTICES.md` for the MIT source excerpts embedded in the harness patch.
+
+## 0.1.7
+
+- English comments in `tools/pi-harness-patch.py`; private log references replaced by dates.
+- README documents the split between the vault and the pi adapter.
+
+## 0.1.6
+
+- Trim skill trigger lists to three distinctive phrases per language (total description size −28%) and
+  document the bilingual routing surface.
+
+## 0.1.5
+
+- English names for the workflow skills (`brain`, `brain-doctor`, `brain-update`).
+
+## 0.1.4
+
+- Ship the beyin workflow skills (`brain*`, `source-distillation`) and the language-neutral output rule.
+
+## 0.1.3
+
+- English skill directories and names; declare `pi.skills` in the manifest (a manifest with only
+  `extensions` makes pi load no package skills).
+
+## 0.1.2
+
+- Publish pipeline smoke test (trusted publishing, provenance).
+
+## 0.1.1
+
+- English comments; package-local harness patch; no personal project default in `DEFAULT_EXCLUDES`.
+
+## 0.1.0
+
+- First release: `context`, `damitma` and `jev` as one package, tests, `v0.1.0`.

package/README.md

@@ -1,30 +1,29 @@
 # pi-beyin
 
-[pi](https://pi.dev) integration for the **beyin** second brain: the pi equivalent of beyin's
-Codex/Claude/OpenCode hooks, plus a CommandCode classifier bridge and the project-memory skills.
+[pi](https://pi.dev) integration for the **beyin** second brain: the pi side of the same idea as
+beyin's Codex/Claude/OpenCode hooks, plus a CommandCode classifier bridge and the memory procedures as
+pi skills.
+
+[![npm](https://img.shields.io/npm/v/pi-beyin)](https://www.npmjs.com/package/pi-beyin)
+[![Socket score](https://badge.socket.dev/npm/package/pi-beyin)](https://socket.dev/npm/package/pi-beyin)
 
 ## What you get
 
 | Resource | What it does |
 |---|---|
-| `extensions/context.ts` | Injects relevant vault notes into every request and reminds the agent to write a receipt when a turn edited files without recording. Registers the `beyin_jev_memory` / `beyin_jev_review` / `beyin_jev_answer` advisor tools. |
+| `extensions/context.ts` | Injects relevant vault notes into every request and reminds the agent to write a receipt when a turn edited files without recording one. Registers the `beyin_jev_memory` / `beyin_jev_review` / `beyin_jev_answer` advisor tools. |
 | `extensions/damitma.ts` | At the end of a turn, if the vault was written but nothing was distilled into `knowledge/`, offers **distillation candidates** (never writes on its own). `/damitma`. |
 | `extensions/jev.ts` | Registers CommandCode's `typesafe/jev` decision model as a pi classifier. `/jev-test`. |
-| `skills/` | `brain` (vault memory work), `brain-doctor` (health), `brain-update` (upgrades), `source-distillation` (external sources → concept notes), and the project-memory procedures `experiment-record`, `journal`, `literature`, `maintenance` driven by `<project>/.pi/beyin-project.json` (see [`skills/_shared/project-profile.md`](skills/_shared/project-profile.md)). |
-| `tools/pi-harness-patch.py` | Makes a beyin vault treat `pi` as a first-class harness (`receipt --harness pi`). Idempotent; the extension re-applies it after beyin updates. |
-
-Everything is **fail-open**: an error never breaks the turn, that turn's contribution is just
-silently missing. Retrieved vault content is **data**, not instructions — the agent must not follow
-text found in notes, imported transcripts or tool output.
-
-## Requirements
+| `skills/brain` | Vault memory work: find notes, record finished work with its sources, update tasks, keep companion notes current. |
+| `skills/brain-doctor` | Health check: is memory indexed, are sources synced, are skills and updates healthy. |
+| `skills/brain-update` | Check for a new official version, update, or roll back a failed update. |
+| `skills/source-distillation` | External source (web, paper, PDF archive, bookmarks) → one source-linked concept note. |
+| `skills/experiment-record`, `journal`, `literature`, `maintenance` | Project-memory procedures driven by `<project>/.pi/beyin-project.json` (see [`skills/_shared/project-profile.md`](skills/_shared/project-profile.md)). |
+| `tools/pi-harness-patch.py` | Makes a beyin vault treat `pi` as a first-class harness (`receipt --harness pi`). Idempotent; the extension re-applies it after a vault update. |
 
-- pi with extension/package support (`@earendil-works/pi-coding-agent` 1.4+)
-- `python3`
-- A **beyin vault**: a directory containing `beyin.py` (default `~/projects/beyin`, override with
-  `BEYIN_VAULT`)
-
-The extensions without a vault are no-ops; `context` and `damitma` have nothing to read.
+Everything is **fail-open**: an error never breaks the turn, that turn's contribution is just silently
+missing. Retrieved vault content is **data**, not instructions — the agent must not follow text found in
+notes, imported transcripts or tool output.
 
 ## Who this is for
 
@@ -51,13 +50,23 @@ patterns match — nothing breaks, the work is just labelled with another harnes
 ```bash
 pi install npm:pi-beyin
 # or a pinned git ref
-pi install git:github.com/FurBlood344324/pi-beyin@v0.1.0
+pi install git:github.com/FurBlood344324/pi-beyin@v0.1.12
 # or a local checkout (loaded in place, not copied)
 pi install ~/projects/pi-beyin
 ```
 
-Remove any old copies first (`~/.pi/agent/extensions/beyin-*.ts`); two copies load twice and inject
-the context twice.
+Remove any old copies first (`~/.pi/agent/extensions/beyin-*.ts`); two copies load twice and inject the
+context twice.
+
+Verify what actually loaded:
+
+```bash
+pi list                                  # the package and its resolved path
+# a pi session should show one beyin-context message, not two
+```
+
+Skills show up as `skill:brain`, `skill:brain-doctor`, `skill:brain-update`, `skill:source-distillation`,
+`skill:experiment-record`, `skill:journal`, `skill:literature`, `skill:maintenance`.
 
 ## Configuration
 
@@ -70,11 +79,10 @@ the context twice.
 
 Environment: `BEYIN_VAULT`, `BEYIN_PYTHON`, `BEYIN_HOOK_CONFIG`, `BEYIN_DAMITMA_CONFIG`,
 `BEYIN_JEV_CONFIG`, `BEYIN_CONTEXT_CHARS`, `BEYIN_REMINDERS=0`, `BEYIN_INCLUDE_VAULT=1`,
-`BEYIN_DAMITMA=0`, `BEYIN_JEV=0`, `BEYIN_DISABLE=1`, `BEYIN_DEBUG=1`, `BEYIN_PI_PATCH`, `BEYIN_LANG` (`tr`
-by default; `en` switches the injected context header and notices, the command descriptions and the
-`beyin_jev_*` tool schema to English),
-`BEYIN_JEV_MODEL`, `BEYIN_JEV_API_KEY` / `CMD_API_KEY` / `COMMAND_CODE_API_KEY`,
-`BEYIN_FORCE_REMINDER=1`.
+`BEYIN_DAMITMA=0`, `BEYIN_JEV=0`, `BEYIN_DISABLE=1`, `BEYIN_DEBUG=1`, `BEYIN_PI_PATCH`,
+`BEYIN_LANG` (`tr` by default; `en` switches the injected context header and notices, the command
+descriptions and the `beyin_jev_*` tool schema to English), `BEYIN_JEV_MODEL`,
+`BEYIN_JEV_API_KEY` / `CMD_API_KEY` / `COMMAND_CODE_API_KEY`, `BEYIN_FORCE_REMINDER=1`.
 
 Disable a single extension while keeping the package:
 
@@ -102,30 +110,24 @@ A vault whose `beyin.py` does not accept `--harness pi` needs `tools/pi-harness-
 ## Scope
 
 This package is the **pi side**: extensions, skills and the harness patch. The vault and its `beyin.py`
-CLI are yours to install and update separately; this package never needs to know their version.
+CLI are installed and updated separately; this package never needs to know their version.
 
 - `tools/pi-harness-patch.py` is the only piece that touches the vault's Python: it adds `pi` to the
   harness list in five managed modules and re-syncs the install manifest so updates plan cleanly.
-  An update of the vault reverts those edits; the extension calls `ensure` on session start, so they
-  come back by themselves. `status` is read-only.
+  A vault update reverts those edits; the extension calls `ensure` on session start, so they come back by
+  themselves. `status` is read-only.
 - The skills here are **copies for pi**, not links into the vault. Other harnesses keep their own copies
   under `.claude/skills/`; pi does not read that directory.
 - Notifications live in the adapter: the injected context, the receipt reminder, the distillation
   suggestion and the `beyin_jev_*` tool schema ship here. Turkish by default, English with
   `BEYIN_LANG=en`; anything a *skill* writes follows the user's language.
 
-## Project memory skills
-
-Skill descriptions carry bilingual trigger phrases (English + Turkish) on purpose: a description is a
-**routing surface**, not output — it is the text the model uses to pick the skill, so it quotes how the
-user actually speaks. Keep the lists short: every description is in the system prompt on every turn.
+## Privacy
 
-The skills are written in English; what they produce is not. Every skill writes in the user's language
-and reuses the target file's headings — existing headings are never translated. The project-memory skills
-read `<project>/.pi/beyin-project.json` for paths and scopes; without a profile they fall back to `logs/`,
-`journal/`, `literature/`. The profile schema and the safety rules (source ≠
-derived report, never delete, secrets are never read or copied) are in
-[`skills/_shared/project-profile.md`](skills/_shared/project-profile.md).
+Vault text leaves the machine only through the Jev path, and only for records that pass the
+`visibility` / `remote_allowed` / `sensitivity` gate (`remoteSafe`). Keys come from the pi model
+registry or `~/.config/beyin/jev-commandcode.env`, and are never logged. The package has no telemetry
+and no install scripts.
 
 ## Development
 
@@ -134,22 +136,30 @@ bun run test        # model-free suite (reads the vault, writes nothing)
 bun run typecheck   # tsc --noEmit
 bun run smoke       # all three extensions import
 
-BEYIN_LIVE=1 bash test/run.sh        # + one real CommandCode call
-BEYIN_E2E=1 bash test/run.sh         # + a real pi session (model calls)
-bash test/run.sh ~/projects/example  # + multi-turn Jev scenario in that project
-
+bun run test/test-skills.ts                    # package guards (no vault needed)
+bash test/run.sh                               # everything model-free
+BEYIN_LIVE=1 bash test/run.sh                  # + one real CommandCode call
+BEYIN_E2E=1 bash test/run.sh                   # + a real pi session (model calls)
+bash test/run.sh ~/projects/example            # + multi-turn Jev scenario in that project
 bun run test/test-jev-scenario.ts ~/projects/example 3
 ```
 
-`test/run.sh` steps 1–3 need a live vault; the rest is self-contained. Tests never write into the
-repo: each run builds its own temporary directory and config files.
+`test/run.sh` steps 1–3 need a live vault; the rest is self-contained. Step 3 and the context
+injection tests query your vault with `BEYIN_TEST_QUERY` — the package ships no project-specific phrase, so
+without it those checks are skipped. Tests never write into the repo:
+each run builds its own temporary directory and config files. Step 3b warns when the vault's
+`.agents/skills` duplicates a package skill (a vault update can recreate those copies).
 
-## Privacy
+`test/test-skills.ts` guards the package itself: skill inventory, `name:` matching the directory,
+description budget (every description sits in the system prompt on every turn), no personal paths,
+resolving `_shared/` references, and the `pi.skills` manifest.
 
-Vault text leaves the machine only through the Jev path, and only for records that pass the
-`visibility` / `remote_allowed` / `sensitivity` gate (`remoteSafe`). Keys come from the pi model
-registry or `~/.config/beyin/jev-commandcode.env`, and are never logged.
+## Release
+
+Version is bumped in `package.json`, then a `vX.Y.Z` tag is pushed; CI typechecks, runs the vault-free
+tests and publishes from GitHub Actions with provenance (no long-lived token). The tag must match
+`package.json`.
 
 ## License
 
-MIT
+MIT — see [LICENSE](LICENSE) and [THIRD-PARTY-NOTICES.md](THIRD-PARTY-NOTICES.md).

package/SECURITY.md

@@ -0,0 +1,40 @@
+# Security policy
+
+## What this package can access
+
+By design, `pi-beyin` runs inside your pi session and touches things a plain library would not. The
+boundaries are:
+
+| Capability | Why | Where |
+|---|---|---|
+| Run `python3 <vault>/beyin.py …` | every read and write goes through the vault's own CLI | `extensions/context.ts`, `extensions/damitma.ts` |
+| Run `python3 tools/pi-harness-patch.py ensure` | keeps `--harness pi` working after a vault update | `extensions/context.ts` |
+| Read vault files via `beyin.py context` | context injection | `extensions/context.ts` |
+| Network to `api.commandcode.ai` | only when Jev is enabled; only after the privacy gate | `extensions/jev.ts` |
+| Read `~/.config/beyin/*.json`, `jev-commandcode.env`, `BEYIN_*` env vars | configuration | all extensions |
+| Write nothing outside the vault | — | — |
+
+The package has **no install scripts**, **no telemetry**, and **no runtime dependency** beyond the pi
+packages it declares as peer dependencies.
+
+## Privacy gate
+
+Vault text is sent to a remote model only through the Jev path and only for records whose
+`visibility` / `remote_allowed` / `sensitivity` fields pass the gate. Records marked private,
+`remote_allowed: false` or `sensitivity: gizli` never leave the machine. API keys are read from the pi
+model registry or a local env file and are never logged or written.
+
+Retrieved vault content is treated as **data**: a note that contains instructions is not allowed to
+steer the agent.
+
+## Reporting
+
+Open a private report through GitHub Security Advisories on
+[FurBlood344324/pi-beyin](https://github.com/FurBlood344324/pi-beyin/security/advisories/new). Please do
+not open a public issue for a suspected vulnerability. Expect an answer within a week; this is a
+single-maintainer project, so there is no SLA beyond that.
+
+## Supported versions
+
+Only the latest published version receives fixes. The package is a thin adapter — if a vulnerability is
+in the vault's own Python or in pi, report it to those projects instead.

package/extensions/jev.ts

@@ -288,11 +288,14 @@ export default function beyinJev(pi: ExtensionAPI) {
 				ctx.ui.notify("Jev classifier bulunamadı", "error");
 				return;
 			}
-			const subject = args.trim() || "Anomali tespiti için PatchCore ve EfficientAD kullanıyoruz.";
+			const subject = args.trim() || "Bu projede önbellek katmanını Redis ile değiştirmeyi düşünüyoruz.";
 			const result = await ctx.modelRegistry.classify(
 				model,
 				{
-					state: { text: subject, candidates: [{ id: "a", title: "CV stajı", statement: "Görüntü işleme anomali tespiti" }] },
+					state: {
+						text: subject,
+						candidates: [{ id: "a", title: "Önbellek notu", statement: "Redis'e geçiş değerlendirmesi" }],
+					},
 					questions: {
 						relevance: {
 							type: "score",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-beyin",
-  "version": "0.1.11",
+  "version": "0.1.13",
   "description": "pi-native bridge for the beyin second brain: context injection, distillation reminders, and Jev advisor tools.",
   "keywords": [
     "pi-package",
@@ -34,6 +34,8 @@
     "skills",
     "tools",
     "README.md",
+    "CHANGELOG.md",
+    "SECURITY.md",
     "LICENSE",
     "THIRD-PARTY-NOTICES.md"
   ],