What is dsh-auto-review AGENTS.md?

A repository instruction file for coding agents working on a DeepSeek Harness plugin. It documents the project layout, coding rules, build steps, and documentation requirements.

Use it to guide edits, builds, reviews, and documentation work in the dsh-auto-review plugin repository.

It gives coding agents the local rules they need to make changes that fit the plugin contract and project structure.

dsh-auto-review AGENTS.md — Instructions file for Codex and OpenCode by PerryLink: A repository instruction file for coding agents working o

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 instructions/perrylink/dsh-auto-review/agents-md

Clone the repo

git clone --depth 1 https://github.com/PerryLink/dsh-auto-review

Made for: Codex, OpenCode.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/perrylink/dsh-auto-review/agents-md.svg)](https://agentmods.dev/instructions/perrylink/dsh-auto-review/agents-md)

Your own site

<a href="https://agentmods.dev/instructions/perrylink/dsh-auto-review/agents-md"><img src="https://agentmods.dev/badge/instructions/perrylink/dsh-auto-review/agents-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 2,216 This file is loaded in full into every session.

When invoked 2,216 The same file — it is already loaded in full.

Security scan A 0 findings. 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.

Model	Per session	Once invoked
Fable 5.1	$0.02216	$0.02216
Opus 5	$0.01108	$0.01108
Sonnet 5	$0.00443	$0.00443
Haiku 4.5	$0.00222	$0.00222

Measured today against content hash 10625530a3a9, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

dsh-auto-review AGENTS.md scanned grade A with 0 findings 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 today.

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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Standalone DeepSeek Harness plugin repository (dsh-auto-review). Development follows the dsh-plugin-guide skill and the official plugin contract; this file records repo-local decisions.

Layout

src/index.ts — function-plugin contract (name/inject/Config/apply; NO default export — the Loader unwraps exports.default ?? exports).
src/config.ts — Schemastery schema + explicit resolveConfig (no hidden ?? default in run() paths).
src/runtime.ts — approval/request answerer, tools/post-execute deny-reason injection, /auto-review command.
src/review.ts — reviewer subagent orchestration, prompt, sanitization, verdict parsing.
src/isolation.ts — the reviewer child's context firewall: the ReviewerChildren registry (prompt announcement before subagents.start, id latch afterwards — the child's first step races the start's resolution) and the agent/pre-step allow-list over message SOURCES that keeps injected context (workspace instruction files, the loop's runtime-context snapshot, third-party injections) out of the reviewer's steps.
src/events.ts — autoReview/state + autoReview/verdict + autoReview/circuit + autoReview/override + autoReview/rejection SessionEventMap members (declaration merging), pure folds, the StateAppend/VerdictAppend/CircuitAppend/OverrideAppend/RejectionAppend surfaces that request the envelope's ignorable: true marker, and the marker-free plain*ResultText builders for audit-disabled hosts.
src/call-id.ts — dual-ruler call-id brand (ToolExecution['callId'] from dsh-tools): host master renamed the dsh-llm CallId brand to ToolCallId, so the package derives the brand locally instead of naming either line's brand.
src/audit.ts — host ignorable-marker capability detection (isMarkedAuditEvent, isUnmarkedHostVersion, peerSessionVersion), the shared seam every audit append gates on.
src/messages.ts — /auto-review command strings in en/zh (the language config selects the table).
src/projection.ts + src/projection-types.ts — the autoReview session-projection unit (host fold + wire schema) and its pure-type outlet (zero value imports, so client programs never drag the host chain). Registered in apply whenever the host provides the session-projection capability (feature-detected: the answerer must work without it).
src/client/ — browser half: the session-header review panel (ReviewPanel.tsx), locale dictionaries, scoped stylesheet, and the client-plugin entry (contract: name = package name, inject, apply(ctx); the bundle follows the shell's window.__ModuleLoader__.load handshake).
src/invariant.ts — invariant companion, exported as dsh-auto-review/invariant. Shipped commented-out in the bundle patch: it needs the invariants service, which spine compositions (headless/ACP) provide but the plain web profile does not.
test/ — vitest; real Context + real Session/ApprovalService/InvariantRegistry from the pinned 0.1.2-rc.1 dev peers, scripted subagent/commands/tools mocks.
fixtures/ — replayable session logs (invariant specs) + config examples.

Read the full file on GitHub · 58 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.

today Changed · +2 tokens per session 10625530a3a9
yesterday Changed · +2 lines · +183 tokens per session 16ca98cc96e0
5d ago First seen · 56 lines · 2,031 tokens per session scan A 4189b350c93f

Subscribe to this mod's changes

dsh-auto-review AGENTS.md is an instructions file published in the GitHub repository PerryLink/dsh-auto-review (134 stars, last pushed today), licensed Apache-2.0. It adds 2,216 tokens to every session, about $0.0111 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

AI Safety Building agents approval cordis deepseek deepseek-harness dsh

dsh-auto-review AGENTS.md

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

AGENTS.md

Layout

What this file has done since we first saw it

Other instructions, from other repositories

dsh-mcp-panel AGENTS.md

dsh-web AGENTS.md

DSH-better-sidebar AGENTS.md

dsh-worktable AGENTS.md

modsearch AGENTS.md

deepseek-harness-desktop AGENTS.md