deepseek

An adapter that runs DeepSeek’s open-source developer-preview agent harness, called dsh, as an OpenAgents workspace agent.

In plain words
What is it for?
Use it to run DeepSeek harness tasks from OpenAgents, with each workspace message handled by a separate dsh process.
Why use it?
It connects the harness to workspace messages while checking its required Node version, exact harness version, entry point, and API key before starting.

Agent

Install

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.

agentmods
npx agentmods add agents/openagents-org/openagents/deepseek
Clone the repo
git clone --depth 1 https://github.com/openagents-org/openagents
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,732 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.01732
Opus 5 $0.00000 $0.00866
Sonnet 5 $0.00000 $0.00346
Haiku 4.5 $0.00000 $0.00173

Measured 2d ago against content hash b1d48c7f2f8c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

deepseek scanned grade A 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 2d ago.

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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

workspace access goes through the documented HTTP API with `curl`, as it does
docs/agents/deepseek.md · 149 lines

How it starts

The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.

DeepSeek Harness (dsh)

DeepSeek's open-source agent harness run as an OpenAgents agent. OpenAgents drives its headless profile — one dsh process per workspace message — and never uses an SDK.

Preview release. DeepSeek ships the harness as a developer preview and says compatibility can break between previews. OpenAgents therefore pins one exact version (0.1.0-rc.6) and refuses to run against any other, including newer ones. See Version pinning.


What the adapter does

dsh --profile headless --patch <private patch> "Read the task file at … and complete the task described in it."
  1. The daemon runs preflight() before joining the workspace: it resolves the dsh entry point, checks the Node version, checks the harness version exactly, and checks that an API key is present. A failure here means the agent never joins, with a precise reason instead of a poll loop that cannot work.
  2. On the first message the adapter composes the headless profile once (--dump-config), so concurrent channels do not race to initialise it.
  3. For each message it collects a bounded channel recap, writes the whole prompt to a private task file, and starts one dsh process.
  4. On exit 0 the process's stdout is the reply. On any other exit the stdout is discarded and stderr is classified into an actionable category.

Behaviour you should expect

No streaming The harness prints only the final assistant message. The workspace shows one "working…" status and then the answer — no intermediate progress.
No conversation memory Each message is a brand-new harness session. Continuity comes from a bounded recap of the channel that OpenAgents injects into every task.
One task at a time per channel Standard BaseAdapter queuing.
60-minute ceiling A run is terminated after 60 minutes total. This is a total budget, not an idle timeout: a long, silent run is normal here and is not killed.

Read the full file on GitHub · 149 lines

Changes

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.

  1. 2d ago First seen · 149 lines · 0 tokens per session scan A b1d48c7f2f8c

Subscribe to this mod's changes

deepseek is an agent published in the GitHub repository openagents-org/openagents (4,020 stars, last pushed 2d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 1,732 tokens. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.