Skip to content
Skills Library
AI & agents · Tool integration

settings-guide

/settings-guide

Help users configure AI endpoints and application settings. Use when the user asks about setting up AI, adding endpoints, configuring models, API keys, proxy settings, user habits/profile preferences, or troubleshooting AI connection issues.

Author
HG-ha ↗
Repository
Argusai ↗
Licence
GPL-3.0
Last repository change
12.11.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 12.11.2026, without references to retired models or APIs.

Licence

GPL-3.0, otwarta.

What it does and when to use it

Help users configure AI endpoints and application settings. Use when the user asks about setting up AI, adding endpoints, configuring models, API keys, proxy settings, user habits/profile preferences, or troubleshooting AI connection issues.

Original author description

Help users configure AI endpoints and application settings. Use when the user asks about setting up AI, adding endpoints, configuring models, API keys, proxy settings, user habits/profile preferences, or troubleshooting AI connection issues.

How it works

  1. Settings Guide
Original SKILL.md headings · 45 lines, 585 words
SKILL.md preview
---
name: settings-guide
description: Help users configure AI endpoints and application settings. Use when the user asks about setting up AI, adding endpoints, configuring models, API keys, proxy settings, user habits/profile preferences, or troubleshooting AI connection issues.
trigger-keywords:
  - "\\b(settings|endpoint|api key|model config|ai connection)\\b"
  - "端点|模型配置|API.?密钥|连接测试|设置页"
---

## Settings Guide

### AI Endpoints
Add endpoints in Settings — two types available:
- **OpenAI Compatible**: works with OpenAI, DeepSeek, Ollama, vLLM, Qwen, and any `/v1/chat/completions` endpoint.
- **Anthropic**: for Anthropic-compatible endpoints.

User fills: base URL, API key, model name. Each endpoint has a priority — higher tried first.

### Connection Issues
- Check base URL reachable, API key valid, model name exists, no firewall/proxy block.

### Proxy Config
Default `127.0.0.1:8866`. Change port if in use, host to `0.0.0.0` for all interfaces.

### Browser
- **Browser binary** card: pick which Chromium-family browser the managed headless instance should use. Empty channel + empty executable_path = auto-detect (system Chrome / Edge / Brave first, Playwright bundle as fallback). Use the "Rescan" button to refresh detection and "Download Playwright Chromium" if no system browser is installed.
- **Reuse my running browser (CDP attach)** card: connect the assistant to the user's real running browser so it can see open tabs and reuse the logged-in session. Two paths:
  - "Relaunch & attach": closes the current Edge/Chrome (warn the user to save unsaved work) and restarts it with `--remote-debugging-port` + the user's default profile.
  - "Connect to endpoint": when the user already started the browser themselves with `--remote-debugging-port=9222`, just enter the endpoint URL.
  - Use "Detach" when finished. The browser keeps running; only the CDP websocket is released.
- For the full workflow, decision rules, and safety guardrails see `browser-attach-workflow`.

### Reverse toolchain (JADX / Frida / APKTool / capa / …)
- **Settings → Reverse toolchain paths**: optional absolute paths when tools are not on PATH. **Scan PATH** detects installed binaries; **Install** downloads supported tools into `data_root/tools/reverse/` and saves paths (needs network + user click).
- With user confirmation, the Agent may call **`reverse_toolchain_install(tool=…)`** for the same managed installs (frida, jadx, apktool, adb, capa, yara, ghidra, innoextract, zipalign, apksigner, objection). **class-dump / jarsigner** remain manual-only — point to docs.
- When a reverse tool returns `blocked: true` / `toolchain_required` and `can_install: true`, prefer managed install over ad-hoc shell/pip. **Do not** `run_shell_snippet` or `sandbox_pip_install` for reverse CLIs.
- Runtime exceptions are different: **managed Python/Node/webcrack/WABT** install from Settings → Sandbox runtimes (or `runtime_install`) and follow `session-sandbox-workflow` / the WASM workflow; bundled **sqlmap** follows the SQLi workflow — none of them belong to the reverse toolchain list above.

### User Habits Profile
- The user habits profile lives in `settings.json` under `user_habits`; it is separate from the knowledge base and from session todos.
- Non-empty profile fields are injected into the assistant system prompt each turn through `profile_prompt_summary`, so the AI should follow them without calling a read tool first.
- The AI should automatically summarize durable user preferences into update suggestions. Examples: preferred response language/style, desired report format, risk tolerance, audit focus, workflows to avoid, or repeated corrections such as "do not do X".
- Do not store one-off task facts, credentials, secrets, target-specific vulnerability details, or temporary instructions in the profile.
- The Settings page section is collapsed by default. Users can edit it manually there.
- Tool governance:
  - `user_profile_read` is read-only.
  - `user_profile_suggest_update` proposes a patch and goes through a confirmation card; use it proactively when a durable preference is observed.
  - `user_profile_apply_update` writes settings and must go through the confirmation framework; the runtime injects `confirm_label=APPLY_USER_PROFILE` only after approval.
  - Do not ask the model to provide `confirm_label` manually. Do not treat global auto-allow as permission to bypass profile-write confirmation.

### Persistence
Settings live in the user-level `settings.json` managed by `PathManager`; they survive restarts. Call `get_settings` before changing.

Versions

The database has no other skills named settings-guide.

Database snapshot: 13.11.2026 · package version a4f3ea46da79 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 95settings-guide-skill
    Triggers when the user asks about settings, preferences, API keys, model switching, theme / language / shortcuts / proxy / backup, or related failures (key invalid, model not found) within the OpenLoaf desktop product — where to configure and how to configure them. **Not for**:…
  • 100manifest
    Install and configure the Manifest observability plugin for your agents. Use when setting up telemetry, configuring API keys or endpoints, troubleshooting plugin connection issues, or verifying the observability pipeline is running.
  • 90dsh-configuration-guide
    Use when configuring DeepSeek Harness profiles, bundles, models, permissions, or workspace preferences.
  • 100manifest
    Install and configure the Manifest observability plugin for your agents. Use when setting up telemetry, configuring API keys, or troubleshooting the plugin.
  • 100manifest
    Install and configure the Manifest observability plugin for your agents. Use when setting up telemetry, configuring API keys, or troubleshooting the plugin.
  • 100manifest
    Install and configure the Manifest observability plugin for your agents. Use when setting up telemetry, configuring API keys, or troubleshooting the plugin.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.