GanyuanRan/Aegis
Software-engineering method pack for coding agents, with skills for baseline-first planning, systematic debugging, prompt hygiene, verification before completion, and repair/retirement tracking.
Install
via GitHub · GitHub source
Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/web/pnpm-workspace.yaml.
What happened when we ran it
Installed cleanly when we ran it
Every command here is run in a throwaway container against a clean profile, and the result is whatever the harness recorded — not a guess from the source. Last run 7d ago.
Show it in your README
For maintainers: the badge serves this listing's latest sandbox verdict, so a re-run updates it on its own — and it links readers to the full result here.
Due diligence
What it is — Aegis is a software engineering method pack for coding agents.
Who it is for — Developers using DeepSeek Harness for complex coding tasks who require baseline-first planning are suitable for installing Aegis. Users who only handle simple requests without needing verification before completion can skip installing it.
Watch out — Sandbox testing passed: it was installed in a new profile and registered by Harness. No obvious issues found.
The verdict — I would install it because it provides verification before completion proofs for coding agents on DeepSeek Harness, which is valuable when proof before "done" is needed.
Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.
Aegis is a DeepSeek Harness plugin that gives coding agents a software-engineering method pack for planning, debugging, prompt hygiene, verification, and repair or retirement tracking.
It projects these methods into a skill-aware host rather than adding a separate coding tool. Before editing, the agent is expected to establish the project's real baseline, including owners, contracts, and boundaries. Before declaring completion, it should produce fresh verification evidence, identify the covered scope, and record residual risk. The pack also keeps retired fallbacks and old paths tied to a retirement trigger instead of leaving them as untracked technical debt. Simple tasks can remain on a fast path; the documented method is intended to add ceremony when the task needs it.
Aegis is aimed at coding-agent workflows where changes need explicit planning and evidence. It is less suitable when the host cannot load skills or when its existing workflow already provides the same planning and verification discipline. The README documents host-specific activation requirements, including native plugin installation for DSH and additional discovery checks where a host uses a separate skill directory.
The detected terminal surface matters: the plugin's verification flow runs commands such as python scripts/aegis-doctor.py --write-config --json from the method-pack root and may pass discovery-root and name-prefix options. That gives the agent a terminal-based path to inspect installation and workspace support, with the agent's available permissions.
Aegis is presented as a method pack projected into a skill-aware coding host. Its documented practices cover baseline-first planning, systematic debugging, prompt hygiene, verification before completion, and two-track repair or retirement tracking. The baseline is expected to include project owners, contracts, and boundaries. Completion evidence should be fresh, state the covered scope, and identify residual risk. Trivial requests can remain on the fast path.
The README requires verification from the installed method-pack root, not from the target project directory. Locate <aegis-method-pack-root> first, then run:
cd <aegis-method-pack-root> && python scripts/aegis-doctor.py --write-config --json
The installation is considered complete only when the JSON contains all of the following:
| Field | Required value |
|---|---|
ok |
true |
workspaceSupport |
available |
configStatus |
configured |
The --write-config option writes the configuration used by the doctor flow, while --json requests machine-readable output.
For the official DeepSeek Harness (dsh), the README specifies native profile-plugin installation with dsh plugin --profile <profile> add github:GanyuanRan/Aegis. The direct-child compatibility path should not be substituted unless the plugin manager is unavailable and compatibility mode is explicitly approved. Host activation and automatic-entry checks must also be completed; file discovery or a generic doctor result is not sufficient when the host defines a plugin, hook, or session-start bootstrap contract.
When a host uses a separate skill discovery directory, repeat verification with --discovery-root <path>. If its host guide defines a skill-directory name prefix, also pass --discovery-name-prefix <prefix>. The README names Codex, Claude Code, OpenCode, and Kimi as other skill-aware hosts.
Written from the project's own documentation and kept in sync with it. Where the two disagree, the source is authoritative — read the README on GitHub
Same category
zhu1090093659
Skill center for the dsh web GUI: browse all loaded skills grouped by source, enable or disable model invocation, create new skills, and delete into a recoverable trash.
Installed cleanly when we ran it
chuspeeism
现代化可灵活嵌入的任务面板，支持 Codex、DeepSeek Harness
Installed cleanly when we ran it
Alisa0808
Turn one topic into a finished Vox-style paper-collage explainer/ad video — automated end to end on Atlas Cloud + ffmpeg. An agent skill.
Installed cleanly when we ran it
superdesigndev
Design skill for UI and marketing graphics on the Superdesign canvas: reads the repo for context, extracts its design system, then generates and iterates branchable design drafts, flow pages, and reusable components through the Superdesign CLI.
Installed cleanly when we ran it
jingyunstudio
基于 Jingyun Studio + DeepSeek Harness (DSH) 打造的一站式 AI 商业化桌面客户端
Did not install when we ran it
