Aegis is a method pack that guides coding agents to inspect a project's baseline, make bounded changes, and verify their work with fresh evidence. It is for people using coding-agent hosts who want fewer unverified changes and less unnecessary process. The catalogue add-ons implement this method through skills, instructions, commands, a hook, and a plugin.
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 GanyuanRan/Aegis --skill verification-before-completiongit clone --depth 1 https://github.com/GanyuanRan/AegisWrote 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/ganyuanran/aegis/verification-before-completion)<a href="https://agentmods.dev/skills/ganyuanran/aegis/verification-before-completion"><img src="https://agentmods.dev/badge/skills/ganyuanran/aegis/verification-before-completion.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00036 | $0.01935 |
| Opus 5 | $0.00018 | $0.00967 |
| Sonnet 5 | $0.00007 | $0.00387 |
| Haiku 4.5 | $0.00004 | $0.00194 |
Grade A, and why
verification-before-completion 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 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.
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.
How it starts
The opening of the file, as written. The whole thing — 201 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Execute
Before any success claim, classify destructive permission needs, choose and run a fresh falsifying check, read its complete result/scope, then select L0/L1/L2. If evidence is partial, stale, failing, or narrower than the claim, downgrade; never claim complete first and verify later.
This Method Pack grants no authoritative GateDecision, PolicySnapshot,
evidence sufficiency, requirement acceptance, or completion authority.
Stop Signals
Stop before claiming or advancing when:
- evidence is uncertain, stale, agent-only, or narrower than the claim;
- the next action is commit, push, PR, merge, tag, publish, release, or handoff;
- task/slice completion is being treated as accepted requirement satisfaction;
- governance or retirement lacks repair/retirement evidence;
- retained old logic lacks a retention reason and retirement trigger; or
- complexity closure is unresolved.
Destructive or irreversible work needs scoped permission; warnings or broad assent do not grant it.
Required Evidence Slots
Keep these slots explicit and auditable:
- Evidence action / check performed:
- Result / exit status:
- Covered scope:
- Uncovered scope:
- Residual risk:
- Confidence grade: A | B | C
A: direct target plus relevant regression evidence; no meaningful unknown.B: direct target evidence with bounded residual risk.C: partial evidence only; do not claim full completion.
When tests shape the claim, include target test and related regression evidence. Place the regression test at the seam that exercises the real bug pattern as it occurs at the call site; a seam too shallow to replicate the triggering chain gives false confidence. If no such seam exists, record the missing seam as the finding (the architecture prevents locking the bug down) and lower confidence. If automation is blocked, give reproducible manual steps and lower confidence. Evidence is not completion authority.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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 · +4 lines 1c31f509980a
- 4d ago Changed · +13 lines 34d566a07ed5
- 8d ago First seen · 184 lines · 36 tokens per session scan A de0c33c532cf
verification-before-completion is a skill published in the GitHub repository GanyuanRan/Aegis (1,172 stars, last pushed yesterday), licensed MIT. It adds 36 tokens to every session and 1,935 once invoked, about $0.0002 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.
Other skills, from other repositories
go-testing
Trigger: Go tests, go test coverage, Bubbletea teatest, golden files. Apply focused Go testing patterns.
rulesync
Generates and syncs AI rule configuration files (.cursorrules, CLAUDE.md, copilot-instructions.md) across 20+ coding tools from a single source. Use when syncing AI rules, running rulesync commands, importing or generating rule files, or managing shared AI coding configurations.
autoprompt
Explicit-only useful-first orchestration. Invoke /autoprompt to turn a mission into one executable roadmap, build dependency-safe lanes, and verify the result with independent reviewers. Never infer invocation from ordinary requests. Never resume from leftover artifacts without an explicit resume instruction.
gitnexus
A code-graph analysis add-on for examining an existing codebase, including symbols, call paths, execution flows, and effects across repositories. It can query GitNexus through its command-line or MCP interfaces.
loop
Full execution protocol for MODE: LOOP — the compound-engineering loop: brainstorm → plan → build → review → improve, iterating under defense-in-depth stop conditions with generator/critic separation, durable resumable state, and mandatory compounding learning capture. Loaded on demand by the architect when the loop…
restore-internals-seams-in-finally-blocks-after-each-test
When delegating a task affected by this skill, include.