dsh-agent-approval
A DeepSeek Harness Web UI plugin that routes sandbox escalation requests to an independent approval subagent and records decisions locally.
Install
$ dsh plugin --profile web add /path/to/dsh-agent-approvalPlugin Overview & Capabilities
AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.
Key Capabilities
- Adds an Agent Approval option to the /permission menu and /agent-approval on|off commands.
- Intercepts escalation approval requests before the normal human approval flow when enabled.
- Spawns an isolated, tool-less approval subagent that returns structured decision, risk level, and rationale data.
- Rejects destructive, irreversible, out-of-scope, or mismatched operations; only allows operations judged safe, reversible, task-aligned, and honestly justified.
- Fails closed when the approval agent cannot start, times out, is cancelled, or returns invalid output.
- Supports a configurable approval provider/model and a configurable 30–600 second timeout.
- Provides a Settings audit view with locally persisted recent approval records, including decision, risk level, model, duration, rationale, and tool parameters.
- Restores the previous sandbox and approval settings when Agent Approval is disabled.
Useful For
- Applying automated policy review to sandbox permission escalations in DSH sessions.
- Avoiding repeated human approval prompts while retaining a rejection-by-default failure mode.
- Reviewing recent automated approval decisions and their associated tool parameters.
Who It Fits
- DeepSeek Harness Web UI users.
- Teams or individuals who want an approval-agent layer for sandbox escalations.
- Developers testing DSH permission workflows.
Documented Limitations
- Designed for the DeepSeek Harness Web UI profile.
- Only affects sessions where the Agent Approval mode is enabled; other sessions retain normal approval behavior.
- Approval decisions are made from limited supplied context: workspace path, recent user message, tool name, escalation rationale, and tool parameter JSON.
- The optional permission-menu shield icon requires running npm run patch:glyph because the upstream glyph mapping has no public registration point.
- The local audit history retains the most recent 200 records.
DSH Compatibility
Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.
Security Signals
Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- MoonlitDropOfBlood/dsh-agent-approval
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 0030177 · 2026-09-11
This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.
Repository Activity
- GitHub stars
- 2GitHub stars
- Forks
- 1
- Open issues
- 0
- Last commit
- 2026-09-06
- Last release
- 2026-09-06
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Preview, create, and edit spreadsheets, docs, slides, and canvases in DeepSeek Harness powered by Univer.
View plugin →Compatibility bridge and host ABI enabling unmodified Pi plugins to run natively on DeepSeek Harness.
View plugin →Local-first AI agent runtime and DSH plugin bundle providing sandboxed sessions, MCP tools, audit logs, and session replay.
View plugin →Import and resume chat histories from 18+ AI coding agents into DeepSeek Harness with reverse export and sync.
View plugin →