boomkin 0.13.6 → 0.13.8

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,13 @@
1
1
  # Changelog
2
2
 
3
+ ## 0.13.8 · 2026-11-18
4
+
5
+ - Review native Hermes 0.23.2 and pin it for fresh installs with its official installer checksum. Its installer no longer gives the builder's uid ownership of the uv it unpacks when run as root, and an exported `GIT_DIR` or a git-managed home directory can no longer redirect its rerun stash and reset to another repository. The native compatibility check passes against it, and its OpenAI Codex OAuth provider is unchanged.
6
+
7
+ ## 0.13.7 · 2026-11-17
8
+
9
+ - Pin the released strategy 0.2.2 pack, from the same upstream revision as the ten packs repinned in 0.13.6. It states the retrieval-time rule as the engine enforces it: no observation may be later than the retrieval time. The catalog keeps 17 packs, 81 skills and 63 workflow IDs, and every pin matches a published upstream release tag.
10
+
3
11
  ## 0.13.6 · 2026-11-17
4
12
 
5
13
  - Pin the released LP 0.7.7, Hyperliquid 0.4.7, routing 0.4.5, payments 0.2.4, agent 0.1.5, infra 0.4.7, lending 0.4.6, yield 0.4.4, staking 0.3.3 and data 0.6.5 packs, all from one upstream revision. They correct wrong provider facts and instructions that contradicted each other across skills, and the LP range helper now reports the exact current tick more than 20,000 ticks from zero. The catalog keeps 17 packs, 81 skills and 63 workflow IDs, every repinned pack matches a published upstream release tag, and the workflow guide links the new revision.
package/README.md CHANGED
@@ -25,7 +25,7 @@ Only want the skills inside Claude Code, Codex, Cursor or another agent you alre
25
25
  Use macOS or Linux with [Bun](https://bun.sh) 1.3.12+ and Git. Windows users can use WSL2.
26
26
 
27
27
  ```bash
28
- bun install -g boomkin@0.13.6
28
+ bun install -g boomkin@0.13.8
29
29
  boomkin desk
30
30
  ```
31
31
 
@@ -147,7 +147,7 @@ Fresh onboarding includes all 17 packs. Repeated `--pack` options choose fewer;
147
147
  ## Updates and recovery
148
148
 
149
149
  ```bash
150
- bun install -g boomkin@0.13.6
150
+ bun install -g boomkin@0.13.8
151
151
  boomkin check --directory "$HOME/.boomkin/hermes"
152
152
  boomkin update --directory "$HOME/.boomkin/hermes"
153
153
  boomkin doctor --directory "$HOME/.boomkin/hermes" --live
@@ -235,8 +235,8 @@
235
235
  "id": "defi-strategy-skills",
236
236
  "source": "galleonlabs/crypto-defi-skills",
237
237
  "description": "Frozen daily strategy validation with cost sensitivity, held-out periods and a reproducible benchmark.",
238
- "version": "0.2.1",
239
- "revision": "27b7cc4478f68dc988bb129a66f8a8f884b76aca",
238
+ "version": "0.2.2",
239
+ "revision": "b87c1a515ab46326dac5a24de3d794296aaf915d",
240
240
  "skills": [
241
241
  "galleon-defi-strategy-backtest"
242
242
  ],
package/dist/cli.js CHANGED
@@ -8371,8 +8371,8 @@ var skills_default = {
8371
8371
  id: "defi-strategy-skills",
8372
8372
  source: "galleonlabs/crypto-defi-skills",
8373
8373
  description: "Frozen daily strategy validation with cost sensitivity, held-out periods and a reproducible benchmark.",
8374
- version: "0.2.1",
8375
- revision: "27b7cc4478f68dc988bb129a66f8a8f884b76aca",
8374
+ version: "0.2.2",
8375
+ revision: "b87c1a515ab46326dac5a24de3d794296aaf915d",
8376
8376
  skills: [
8377
8377
  "galleon-defi-strategy-backtest"
8378
8378
  ],
@@ -8410,7 +8410,7 @@ var skills_default = {
8410
8410
  // package.json
8411
8411
  var package_default = {
8412
8412
  name: "boomkin",
8413
- version: "0.13.6",
8413
+ version: "0.13.8",
8414
8414
  description: "Your open-source DeFi desk: native Hermes research projects, verifiable evidence and independent financial skills.",
8415
8415
  type: "module",
8416
8416
  license: "MIT",
@@ -8578,9 +8578,9 @@ async function terminateOwnedProcess(child, group, graceMs = 2000) {
8578
8578
  }
8579
8579
 
8580
8580
  // src/hermes.ts
8581
- var HERMES_VERSION = "0.23.1";
8582
- var HERMES_COMMIT = "3b421578588243f920886e6fe79a6faa87ea026e";
8583
- var HERMES_INSTALLER_SHA256 = "abf338605dae6dd28d101673ac6aed7995c955745db46d47676bc591a7957209";
8581
+ var HERMES_VERSION = "0.23.2";
8582
+ var HERMES_COMMIT = "ae2a81ccc00e9b7c06301523e5bb09608397149c";
8583
+ var HERMES_INSTALLER_SHA256 = "c73dad4dcf7c9ecadb98578e28cc763d23d7be3fa1b513545319038795493214";
8584
8584
  var HERMES_INSTALLER_URL = `https://raw.githubusercontent.com/NousResearch/hermes-agent/${HERMES_COMMIT}/scripts/install.sh`;
8585
8585
  function absent(error) {
8586
8586
  return error.code === "ENOENT";
package/docs/CHATGPT.md CHANGED
@@ -6,7 +6,7 @@ Examples use the installed `boomkin` command; from source, use `bun run boomkin`
6
6
 
7
7
  [OpenAI lists Hermes Agent as a ChatGPT plan-usage partner](https://learn.chatgpt.com/docs/sign-in-with-chatgpt). Account sign-in, permission to consume plan usage, a successful inference and usage attributed to Hermes are separate milestones. Boomkin is an onboarding layer, not a separately registered OpenAI partner.
8
8
 
9
- Boomkin retains reviewed Hermes 0.23.1 at `3b421578588243f920886e6fe79a6faa87ea026e`. Its native `auth add openai-codex --type oauth` supports a fresh ChatGPT device login, profile-native credential pools and refresh; `auth status` reports authentication. Its source uses the Codex OAuth client and Codex inference route. This proves native ChatGPT OAuth capability; it does **not** prove the new Hermes partner plan-consent or per-app attribution path. Use the [official Nous portal](https://portal.nousresearch.com/) for the partner connection and review its actual permission screen. Do not replace the pinned runtime or invent partner client IDs/endpoints to bridge that gap.
9
+ Boomkin retains reviewed Hermes 0.23.2 at `ae2a81ccc00e9b7c06301523e5bb09608397149c`. Its native `auth add openai-codex --type oauth` supports a fresh ChatGPT device login, profile-native credential pools and refresh; `auth status` reports authentication. Its source uses the Codex OAuth client and Codex inference route. This proves native ChatGPT OAuth capability; it does **not** prove the new Hermes partner plan-consent or per-app attribution path. Use the [official Nous portal](https://portal.nousresearch.com/) for the partner connection and review its actual permission screen. Do not replace the pinned runtime or invent partner client IDs/endpoints to bridge that gap.
10
10
 
11
11
  ```sh
12
12
  boomkin chatgpt --action login --directory /absolute/path/to/boomkin-profile
@@ -10,7 +10,7 @@ Provider-specific review dates and limits appear below. The discovery directory
10
10
 
11
11
  `boomkin onboard` runs the official pinned Hermes installer when needed and opens `hermes setup model` with the selected `HERMES_HOME`. Use `boomkin model` to revisit setup, and `boomkin start` for native chat. From a source checkout, use `bun run boomkin`. Model keys, OAuth and local models use Hermes's facilities. A configured model name or CLI version is not proof of authentication.
12
12
 
13
- Hermes v0.23.1 (upstream tag `v0.23.1`) at [the reviewed source](https://github.com/NousResearch/hermes-agent/tree/3b421578588243f920886e6fe79a6faa87ea026e) was tested. Boomkin pins the official installer and runtime for a fresh install. Existing installations are preserved. Native setup supports optional browser, tools and messaging features; configure those through Hermes rather than a second runtime.
13
+ Hermes v0.23.2 (upstream tag `v0.23.2`) at [the reviewed source](https://github.com/NousResearch/hermes-agent/tree/ae2a81ccc00e9b7c06301523e5bb09608397149c) was tested. Boomkin pins the official installer and runtime for a fresh install. Existing installations are preserved. Native setup supports optional browser, tools and messaging features; configure those through Hermes rather than a second runtime.
14
14
 
15
15
  The selected directory holds `config.yaml`, `.env`, `SOUL.md`, `skills/` and native state. Boomkin stores its own records under `.boomkin/`. Commands explicitly select the root profile so a sticky Hermes profile preference cannot redirect them elsewhere. Do not move a configured profile without reviewing saved paths and credentials.
16
16
 
package/docs/DESK.md CHANGED
@@ -3,7 +3,7 @@
3
3
  Open Boomkin in your browser, choose a question and keep the result with its sources. Market snapshots, public Hyperliquid wallet reviews and daily strategy tests work without a model account or funded wallet.
4
4
 
5
5
  ```bash
6
- bun install -g boomkin@0.13.6
6
+ bun install -g boomkin@0.13.8
7
7
  boomkin desk
8
8
  ```
9
9
 
@@ -9,7 +9,7 @@
9
9
  Use macOS or Linux with [Bun](https://bun.sh) 1.3.12+ and Git. Keep Bun's global bin directory on your shell path. Windows users can use WSL2; an existing native Windows Hermes installation can also be reused through its official flow.
10
10
 
11
11
  ```bash
12
- bun install -g boomkin@0.13.6
12
+ bun install -g boomkin@0.13.8
13
13
  boomkin desk
14
14
  ```
15
15
 
@@ -79,7 +79,7 @@ boomkin model --directory "$HOME/research-desk"
79
79
 
80
80
  Hermes owns chat, tools, memory, sessions, scheduling and model authentication. Boomkin supplies reviewed financial procedures, source pins, installation checks and research projects. Other harness adapters install [skills only](HARNESSES.md).
81
81
 
82
- A fresh runtime installation uses reviewed Hermes 0.23.1, upstream tag `v0.23.1`, through its pinned official installer, which manages its own Python 3.14 runtime. A working existing Hermes installation at 0.21.0 or newer is reused; the compatibility check covers 0.23.1. Browser/computer-use dependencies are skipped initially; configure optional features through native Hermes setup. The reviewed runtime source and installer checksum live in [src/hermes.ts](../src/hermes.ts).
82
+ A fresh runtime installation uses reviewed Hermes 0.23.2, upstream tag `v0.23.2`, through its pinned official installer, which manages its own Python 3.14 runtime. A working existing Hermes installation at 0.21.0 or newer is reused; the compatibility check covers 0.23.2. Browser/computer-use dependencies are skipped initially; configure optional features through native Hermes setup. The reviewed runtime source and installer checksum live in [src/hermes.ts](../src/hermes.ts).
83
83
 
84
84
  CoinGecko is the sole onboarding data connection. Optional provider setup stays explicit; [provider discovery](PROVIDERS.md) separates source references from the smaller native connection list. Tool availability, account authority and financial execution are different outcomes.
85
85
 
package/docs/UPDATES.md CHANGED
@@ -9,7 +9,7 @@ Maintain Boomkin, its selected skill packs and the Hermes runtime separately. Us
9
9
  For the installed CLI and default profile:
10
10
 
11
11
  ```bash
12
- bun install -g boomkin@0.13.6
12
+ bun install -g boomkin@0.13.8
13
13
  boomkin check --directory "$HOME/.boomkin/hermes"
14
14
  boomkin update --directory "$HOME/.boomkin/hermes"
15
15
  boomkin doctor --directory "$HOME/.boomkin/hermes" --live
@@ -22,7 +22,7 @@ Replace the directory in every command if you use another profile. Restart an ac
22
22
  | --- | --- |
23
23
  | Boomkin CLI and packaged guides | Install the desired published Boomkin version with Bun |
24
24
  | Selected skill packs | Run `boomkin update` in the chosen profile |
25
- | Hermes runtime | Use native Hermes updates, then rerun `doctor --live`. Boomkin has reviewed 0.23.1; from 0.22, source installs follow stable releases by default |
25
+ | Hermes runtime | Use native Hermes updates, then rerun `doctor --live`. Boomkin has reviewed 0.23.2; from 0.22, source installs follow stable releases by default |
26
26
  | Provider credentials and tools | Follow the provider's native [connection setup](CONNECTIONS.md) |
27
27
 
28
28
  From a source checkout, first pull the reviewed branch and its locked dependencies, then use `bun run boomkin` in place of `boomkin`:
package/docs/WORKFLOWS.md CHANGED
@@ -19,11 +19,11 @@ The [interactive explorer](https://galleonlabs.github.io/boomkin/#workflows) sea
19
19
  | Protocol | Workflow and task ID | Installable skill |
20
20
  | --- | --- | --- |
21
21
  | Hyperliquid | <a id="find-fees-funding-costs-and-exposure-in-a-public-wallet"></a><a id="public-wallet-diagnostics"></a>Find fees, funding costs and exposure in a public wallet<br>[`hyperliquid-wallet-audit`](https://galleonlabs.github.io/boomkin/?workflow=hyperliquid-wallet-audit#workflows) | [`hyperliquid-wallet-audit`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/hyperliquid/skills/hyperliquid-wallet-audit/SKILL.md) |
22
- | Strategy research | <a id="validate-a-frozen-rule-with-higher-costs-and-a-held-out-period"></a><a id="frozen-strategy-validation"></a>Validate a frozen rule with higher costs and a held-out period<br>[`strategy-validation`](https://galleonlabs.github.io/boomkin/?workflow=strategy-validation#workflows) | [`galleon-defi-strategy-backtest`](https://github.com/galleonlabs/crypto-defi-skills/blob/27b7cc4478f68dc988bb129a66f8a8f884b76aca/packages/strategy/skills/galleon-defi-strategy-backtest/SKILL.md) |
22
+ | Strategy research | <a id="validate-a-frozen-rule-with-higher-costs-and-a-held-out-period"></a><a id="frozen-strategy-validation"></a>Validate a frozen rule with higher costs and a held-out period<br>[`strategy-validation`](https://galleonlabs.github.io/boomkin/?workflow=strategy-validation#workflows) | [`galleon-defi-strategy-backtest`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/strategy/skills/galleon-defi-strategy-backtest/SKILL.md) |
23
23
  | Hyperliquid | <a id="research-a-hyperliquid-market-before-planning-a-trade"></a><a id="hyperliquid-research-and-trade-review"></a>Research a Hyperliquid market before planning a trade<br>[`hyperliquid-market`](https://galleonlabs.github.io/boomkin/?workflow=hyperliquid-market#workflows) | [`hyperliquid-analyze`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/hyperliquid/skills/hyperliquid-analyze/SKILL.md) |
24
24
  | Hyperliquid | <a id="review-a-completed-trade-against-its-original-intent"></a>Review a completed trade against its original intent<br>[`hyperliquid-review`](https://galleonlabs.github.io/boomkin/?workflow=hyperliquid-review#workflows) | [`hyperliquid-review`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/hyperliquid/skills/hyperliquid-review/SKILL.md) |
25
25
  | Market data | <a id="capture-a-bounded-market-snapshot-with-source-evidence"></a>Capture a bounded market snapshot with source evidence<br>[`market-snapshot`](https://galleonlabs.github.io/boomkin/?workflow=market-snapshot#workflows) | [`galleon-defi-market-snapshot`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/data/skills/galleon-defi-market-snapshot/SKILL.md) |
26
- | Strategy research | <a id="test-a-strategy-against-a-reproducible-daily-price-history"></a>Test a strategy against a reproducible daily price history<br>[`strategy-backtest`](https://galleonlabs.github.io/boomkin/?workflow=strategy-backtest#workflows) | [`galleon-defi-strategy-backtest`](https://github.com/galleonlabs/crypto-defi-skills/blob/27b7cc4478f68dc988bb129a66f8a8f884b76aca/packages/strategy/skills/galleon-defi-strategy-backtest/SKILL.md) |
26
+ | Strategy research | <a id="test-a-strategy-against-a-reproducible-daily-price-history"></a>Test a strategy against a reproducible daily price history<br>[`strategy-backtest`](https://galleonlabs.github.io/boomkin/?workflow=strategy-backtest#workflows) | [`galleon-defi-strategy-backtest`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/strategy/skills/galleon-defi-strategy-backtest/SKILL.md) |
27
27
  | Research | <a id="review-an-investment-thesis-against-explicit-invalidation-conditions"></a>Review an investment thesis against explicit invalidation conditions<br>[`thesis-review`](https://galleonlabs.github.io/boomkin/?workflow=thesis-review#workflows) | [`galleon-defi-data`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/data/skills/galleon-defi-data/SKILL.md) |
28
28
  | Aave | <a id="stress-test-an-aave-v3-position"></a>Stress-test an Aave V3 position<br>[`aave-health`](https://galleonlabs.github.io/boomkin/?workflow=aave-health#workflows) | [`galleon-aave-position`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/lending/skills/galleon-aave-position/SKILL.md) |
29
29
  | Morpho | <a id="inspect-a-morpho-market-or-vault"></a>Inspect a Morpho market or vault<br>[`morpho-market`](https://galleonlabs.github.io/boomkin/?workflow=morpho-market#workflows) | [`galleon-morpho-market`](https://github.com/galleonlabs/crypto-defi-skills/blob/b87c1a515ab46326dac5a24de3d794296aaf915d/packages/lending/skills/galleon-morpho-market/SKILL.md) |
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "boomkin",
3
- "version": "0.13.6",
3
+ "version": "0.13.8",
4
4
  "description": "Your open-source DeFi desk: native Hermes research projects, verifiable evidence and independent financial skills.",
5
5
  "type": "module",
6
6
  "license": "MIT",