Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add znlgis/my-opencode-deepseek-config --skill opencode-configgit clone --depth 1 https://github.com/znlgis/my-opencode-deepseek-configWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/znlgis/my-opencode-deepseek-config/opencode-config)<a href="https://agentmods.dev/skills/znlgis/my-opencode-deepseek-config/opencode-config"><img src="https://agentmods.dev/badge/skills/znlgis/my-opencode-deepseek-config/opencode-config/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/znlgis/my-opencode-deepseek-config/opencode-config"><img src="https://agentmods.dev/badge/skills/znlgis/my-opencode-deepseek-config/opencode-config.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00080 | $0.03517 |
| Opus 5.5 | $0.00032 | $0.01407 |
| Sonnet 5.5 | $0.00016 | $0.00703 |
| Haiku 4.5 | $0.00008 | $0.00352 |
Grade C, and why
opencode-config scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- `deny` `.env*` reads (except `.env.example`); `ask` on destructive shell (`rm -rf`, `git push -f`, `git reset --hard`, PowerShell/cmd equivalents) and `external_directory`. Cover shell variants (Unix + Windows) so guar How it starts
The opening of the file, as written. The whole thing — 168 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenCode Config Authoring (v2)
For the authoritative v2 shapes, read the schema in the local docs repo rather
than reconstructing them from memory — packages/schema/src/config*.ts is the
loader's own contract:
C:\Users\Administrator\.local\share\opencode\repos\github.com\anomalyco\opencode@dev.
This file only covers this repository's local conventions plus the traps already
paid for.
Repository layout
| Path | Role |
|---|---|
opencode.jsonc |
Global config: model, permissions, plugins, agents, commands, compaction |
AGENTS.md |
Global rules auto-loaded into every agent's context |
agents/<name>.md |
One custom agent per file (frontmatter + system prompt) |
skills/<name>/SKILL.md |
On-demand skills, auto-discovered from the config dir |
Hard constraints
- Only
deepseek/deepseek-v4-proand the natively multimodaldeepseek/deepseek-flash. Never a third model; flash handles both text and visual input. A model variant (deepseek/deepseek-flash#low) is a mode of an allowed model, not a third model. - v2-only, native keys. This repo targets OpenCode v2 (the desktop app's bundled
opencode-cli.exe, measured on 2.1.0). OpenCode 1.18.x is not supported. Write the v2-native spellings —plugins,providers,agents,commands,permissions,media— never the v1 spellings (plugin,provider,agent,command,permission,attachment,small_model,autoupdate, top-levelsubagent_depth). - The
opencodeonPATHis the wrong binary. Measured 2026-10-16: WinGet ships 1.18.4 there, while the runtime in use is the desktop app's 2.1.0. Always invoke"$env:LOCALAPPDATA\Programs\@opencode-aidesktop\resources\opencode-cli.exe";opencode --version→1.18.4means you are about to test the wrong thing.
Target runtime: desktop v2 (CLI 2.1.0 measured)
v2 loads opencode.json/opencode.jsonc through two paths: native decode,
or — if the file contains ANY v1 key — the official V1→V2 migration
(isV1 trigger keys: logLevel, server, command, reference, snapshot,
plugin, autoshare, disabled_providers, enabled_providers, small_model,
mode, agent, provider, permission, tools, attachment, layout).
This repo standardizes on the native path: keep every v1 key out of the file,
so the loader never takes the migration branch.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday Changed a49827950c35
- 5d ago Changed · +31 lines c018b30a852d
- 9d ago Changed · +16 lines bec546e97443
- 14d ago Changed · +12 lines · +4 tokens per session da860e0f152f
- 1mo ago Changed · +16 lines ca8baf9c8a46
- 1mo ago Changed · +21 lines dfec8c4057d1
- 1mo ago Changed cf533ad38706
- 1mo ago Changed fed94c1a750d
- 1mo ago First seen · 72 lines · 76 tokens per session scan C 9883e79b060a
opencode-config is a skill published in the GitHub repository znlgis/my-opencode-deepseek-config (59 stars, last pushed yesterday), licensed MIT. It adds 80 tokens to every session and 3,517 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
my-opencode-deepseek-config
Use when deploying, tuning, or extending the OpenCode × DeepSeek V4 optimal configuration (orchestrator routing, 12 agents, thinking tiers, prompt-cache prefix discipline, limit.input compaction windows, permissions baseline, 23 skills, slash commands) — installing to another machine, adding an agent or skill…
orca-emulator-android
Android device and emulator control from inside Orca over adb, with the live device view in Orca's emulator pane. Use when driving an adb-connected emulator or phone on Windows, Linux, or macOS: booting AVDs, taps, swipes, typing, hardware buttons, rotation, app install and launch, runtime permissions, the…
skill-creator
Create, install, or update skills in the workspace. Use when (1) installing a skill from a URL or remote source, (2) creating a new skill from scratch, (3) updating or restructuring existing skills. Always use this skill for any skill installation or creation task.
ralplan
Consensus planning entrypoint that auto-gates vague ralph/autopilot/team requests before execution.
remember
Review reusable project knowledge and decide what belongs in project memory, notepad, or durable docs.
image-generator
Generate or edit images from text and local files, image URLs, or data URLs through image API. Use when the user asks to create an image, generate from one or more reference images, combine styles or elements across references, or edit a marked region while blending the result into the original scene.