Import AITURK IDE 1.0.0-beta.1 from Hermes 63279301; preserve MIT license

This commit is contained in:
2026-09-05 13:26:46 +03:00
commit 03634b1ca3
11340 changed files with 3442369 additions and 0 deletions
@@ -0,0 +1,147 @@
---
name: actual-setup
description: Set up Actual Computer (actual.inc) inference in Hermes.
version: 2.0.0
author: shl0ms + Hermes Agent
license: MIT
platforms: [linux, macos, windows]
metadata:
hermes:
tags: [actual, actual-inc, provider, local-inference, relay, gguf, setup]
category: devops
---
# Actual Computer Setup Skill
Sets up [actual.inc](https://actual.inc) (Actual Computer) as a Hermes inference
provider. Actual turns the user's own hardware into a private inference cluster
and exposes an OpenAI-compatible API two ways: a hosted end-to-end-encrypted
relay at `https://api.actual.inc` (authenticated with an `ac_` key), and a local
on-device daemon at `http://127.0.0.1:8080` (no auth on loopback). This skill
does not install the Actual daemon for the user — device authorization requires
a human in a browser.
## When to Use
- User wants to add actual.inc as an inference provider (cloud relay or local).
- User has an `ac_` key and wants Hermes routed through their Actual cluster.
- User wants fully-local, on-device inference via the Actual daemon.
- Troubleshooting: Actual requests failing with cryptic 400s or empty streams.
## Prerequisites
- Hermes has **first-class `actual` provider support** (provider id `actual`,
aliases `actual-computer`, `actualcomputer`, `aci`). Do NOT configure Actual
as a `custom_providers` / `providers.actual.*` entry on current Hermes — the
built-in provider owns the name and handles base-url normalization, the
Responses transport, and local no-auth automatically.
- Relay mode: an Actual account and an `ac_` inference key from
https://actual.inc/user/keys.
- Local mode: the user has installed the daemon
(`curl -fsSL "https://actual.inc/install" | bash`) and completed device
authorization by running `actual` once and opening the printed
`https://actual.inc/device?code=...` URL in a browser. Relay that URL to the
user and WAIT — never invent an email or authorize on their behalf. Codes
expire in 5 minutes; re-run `actual` for a fresh one.
## How to Run
### Relay / API mode
1. Put the key in `.env` (secrets only — never config.yaml):
append `ACTUAL_API_KEY=ac_...` to `~/.hermes/.env`.
2. Verify the key and discover models with `terminal`:
```bash
curl -s https://api.actual.inc/v1/models -H "Authorization: Bearer $ACTUAL_API_KEY"
```
3. Select provider + model:
```bash
hermes config set model.provider actual
hermes config set model.default "MODEL_ID_FROM_DISCOVERY"
```
4. Verify end-to-end:
```bash
hermes chat -Q -q "Reply with exactly: ACTUAL_OK" --provider actual -m MODEL_ID
```
### Local mode
1. Human has installed + authorized the daemon (see Prerequisites).
2. Download and load a model (scriptable once authorized):
```bash
actual models search "qwen2.5 0.5b instruct gguf" --limit 8 --no-prompt
# Downloads REQUIRE an explicit quantization (409 ambiguous_model_download otherwise):
actual models download "Qwen/Qwen2.5-0.5B-Instruct-GGUF/Q4_K_M"
actual models list # note the INSTALLED name (differs from download id)
actual models load "qwen2.5-0.5b-instruct-q4_k_m" # load by installed name
```
3. Point Hermes at the daemon. `ACTUAL_BASE_URL` with a loopback host flips the
built-in provider into local no-auth mode automatically — no key needed:
append `ACTUAL_BASE_URL=http://127.0.0.1:8080` to `~/.hermes/.env`, then:
```bash
hermes config set model.provider actual
hermes config set model.default "INSTALLED_MODEL_NAME"
```
4. Verify (reduced toolset — see context-window pitfall below):
```bash
hermes chat -Q -q "Reply with exactly: LOCAL_OK" --provider actual -m INSTALLED_NAME -t file,web
```
## Quick Reference
| Thing | Value |
|---|---|
| Hosted relay | `https://api.actual.inc/v1` (normalized from bare host automatically) |
| Local daemon | `http://127.0.0.1:8080/v1` (no auth on loopback) |
| Key env var | `ACTUAL_API_KEY` (`ac_...`) |
| Base URL env var | `ACTUAL_BASE_URL` (loopback host ⇒ local no-auth mode) |
| Provider id / aliases | `actual` / `actual-computer`, `actualcomputer`, `aci` |
| Transport | Responses API (`codex_responses`) — built-in, do not override |
| Cluster pinning | `X-Cluster-ID` header via `providers.actual.extra_headers` in config.yaml |
| Model size guide | 0.5B Q4_K_M ~470MB (toy), 7-8B Q4_K_M ~4.5GB (daily driver), 32B ~20GB |
## Pitfalls
1. **reasoning_effort trap (handled by Hermes since the first-class provider).**
Actual's SGLang/vLLM backends accept only `none/low/medium/high/max`;
`xhigh`/`ultra` used to fail with a cryptic
`Expecting value: line 1 column 1 (char 0)` (a wrapped HTTP 400). The
built-in provider clamps `xhigh→high` and `ultra→max` on the wire. If a
request still 400s this way on an old Hermes, set a per-model cap:
`agent.reasoning_overrides.<model>: high` in config.yaml.
2. **Context-window overflow on small local models.** Hermes' default toolset
is ~26k tokens of schemas plus a ~9k-token system prompt. A model loaded
with a 32k context overflows before the first turn, and llama.cpp-family
servers emit a bare `data: [DONE]` — Hermes reports
`Provider returned an empty stream with no finish_reason`. This is NOT an
SSE bug. Fixes: restrict tools (`-t file,web`), load the model with a
larger `n_ctx`, or pick a >=64k-context model for the full toolset.
Upstream tracking: #51448 (do not file new issues; add evidence there).
Related but distinct: #65631 (HTTP-200 SSE carrying a 400), #56516
(reasoning-only streams).
3. **Download ids vs installed names.** `actual models download` takes
`repo/QUANT` and 409s without an explicit quantization;
`actual models load` takes the INSTALLED name from `actual models list`.
4. **Reasoning models returning empty content.** GLM/Qwen reasoning variants
emit thinking in a separate `reasoning` field and can burn a small
`max_tokens` entirely on reasoning. Give generous max_tokens before
assuming failure.
5. **Do not create a custom provider named `actual`.** Older setup guides
(pre first-class support) wrote `providers.actual.*` config blocks. On
current Hermes the built-in provider wins the name; stale custom blocks
are ignored or conflict. Remove them and use the env vars + model.provider
flow above.
## Verification
```bash
# Relay:
hermes chat -Q -q "Reply with exactly: ACTUAL_OK" --provider actual -m MODEL
# Local (small model — reduced toolset):
hermes chat -Q -q "Reply with exactly: LOCAL_OK" --provider actual -m MODEL -t file,web
# Provider status (local no-auth shows key_source=local-offline):
hermes status
```
For other OpenAI-compatible clients (e.g. OpenCode), see
`references/opencode.md`.
@@ -0,0 +1,91 @@
# actual.inc as an OpenCode provider
Verified end-to-end 2026-07 (OpenCode 1.18.3, macOS). Adds Actual's relay/GLM
cluster to OpenCode as a custom OpenAI-compatible provider.
## Design: secret in auth.json, config in opencode.json
OpenCode auto-injects a credential when the provider **id** in `opencode.json`
matches a credential **id** in `~/.local/share/opencode/auth.json`. So put the
key in auth.json and NOTHING sensitive goes in opencode.json. This is more robust
than `options.apiKey: "{env:...}"` with the var name, because `{env:...}` only resolves
if the var is exported in the shell OpenCode launches from — and the Actual key
is typically only in `~/.hermes/.env`, not a shell profile, so the env form
breaks outside an inheriting terminal.
### 1. Add the credential to auth.json
File: `~/.local/share/opencode/auth.json`. Shape (preserve existing entries):
```json
{
"anthropic": { "type": "api", "key": "..." },
"actual": { "type": "api", "key": "ac_..." }
}
```
Do this with a read-modify-write (json load, add the `actual` key, dump) so the
other credentials stay intact — don't overwrite the file.
### 2. Add the provider to opencode.json
File: `~/.config/opencode/opencode.json` (or `~/.opencode.json`). Add under
`provider` alongside anything already there. NO `apiKey` field — it comes from
auth.json by id match.
```json
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"actual": {
"npm": "@ai-sdk/openai-compatible",
"name": "Actual (GLM cluster)",
"options": {
"baseURL": "https://api.actual.inc/v1",
"headers": {
"X-Cluster-ID": "<cluster-id-hash>"
}
},
"models": {
"glm-5.2-nvfp4": {
"name": "GLM-5.2 (b300x8)",
"limit": { "context": 1048576, "output": 65536 }
}
}
}
}
}
```
- `npm`: `@ai-sdk/openai-compatible` for `/v1/chat/completions`. Use
`@ai-sdk/openai` only if the model needs `/v1/responses`.
- `options.headers.X-Cluster-ID`: pin to a specific cluster (optional; omit to
let the relay route). Get the hash from the Actual console URL
(`console/computers?cluster=<hash>`).
- `models.<id>`: the id MUST match what `GET /v1/models` returns. Discover it
first: `curl -s https://api.actual.inc/v1/models -H "Authorization: Bearer ac_..." -H "X-Cluster-ID: <hash>"`.
- `limit`: lets OpenCode track remaining context (custom providers don't get
this from models.dev). GLM-5.2 context = 1_048_576.
### 3. Verify live (headless)
```bash
opencode run -m actual/glm-5.2-nvfp4 "Reply with exactly this text: OPENCODE_ACTUAL_OK"
```
OpenCode DOES use the `provider/model` slash form on the CLI (unlike Hermes,
where the slash form 404s custom providers). Expect the exact reply. Run a second
reasoning check (e.g. "What is 17 * 23?") since GLM-5.2 is a reasoning model.
## Why no reasoning_effort trap here
The Actual relay rejects `reasoning_effort: xhigh` with an HTTP 400 (see the
`hermes-custom-providers` skill, pitfall 2). Hermes hits this because it forwards
its global `agent.reasoning_effort`. OpenCode's ai-sdk does NOT send that param,
so Actual + OpenCode works with zero reasoning config. No `reasoning_overrides`
equivalent needed.
## Gotchas
- `auth.json` is the same store `/connect` writes; editing it directly is fine
and equivalent. `opencode auth list` should then show `actual` under
Credentials.
- If discovery/models don't appear, confirm the provider id in opencode.json
EXACTLY matches the auth.json credential id (`actual` == `actual`).
- No git-tracking risk on the default config dir (`~/.config/opencode` is not a
repo), but still keep the key in auth.json, not opencode.json, as the habit.