Observed since 20 Oct, 20:19 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
Control and audit plane CLI for AI coding agents.
Control & audit plane for AI coding agents — one constitution, every agent, safer commits.
Aegisure is a local-first CLI for teams and developers using coding agents such as Codex, Claude Code, Cursor, Copilot, Cline, and Roo. It gives every repo a shared Constitution, scans AI-generated diffs before they land, exports rule files that agents can read, and creates constrained repair prompts when a change is risky.
The static core is LLM-free. It runs locally, works without API keys, and does not send your code to a model.
Aegisure.md Constitution for a repository.AGENTS.md, CLAUDE.md, .cursorrules, .clinerules, and .github/copilot-instructions.md.aegisure commit.aegisure scan --changed --json.pip install aegisure
Requires Python 3.10 or newer.
Initialize Aegisure in a repo:
aegisure init
aegisure export
Stage a risky change:
echo 'OPENAI_API_KEY = "sk-proj-realfakekey1234567890abcdef"' >> app.py
git add app.py
Scan the staged diff:
aegisure scan --staged
Example output:
Aegisure verdict: block (100/100)
1 files changed, 1 additions, 0 deletions, 1 risk findings.
- CRITICAL secret_in_diff at app.py:1: This added line looks like a secret, token, password, private key, or credential.
Create a constrained repair prompt:
aegisure repair --staged --agent codex
aegisure init Generate the canonical Aegisure.md Constitution.
aegisure export Export Aegisure.md into standard agent memory files.
aegisure scan Analyze a local or CI diff using the LLM-free static core.
aegisure repair Generate a constrained repair prompt for risky diffs.
aegisure commit Commit changes and capture declared agent provenance.
aegisure login Store local workspace metadata for CLI use.
aegisure doctor Run a read-only repository readiness check.
aegisure run Start or end a git snapshot-and-scan session.
aegisure rewind Prepare a safe git-based rollback for an Aegisure-tagged commit.
The public CLI is intentionally static-core and LLM-free.
Copy .github/workflows/aegisure.yml.example into your repository as .github/workflows/aegisure.yml:
name: Aegisure
on:
pull_request:
branches: ["**"]
jobs:
scan:
runs-on: ubuntu-latest
permissions:
contents: read
pull-requests: read
steps:
- uses: actions/checkout@v4
with:
fetch-depth: 0
- uses: actions/setup-python@v5
with:
python-version: "3.12"
- run: python -m pip install aegisure
- run: aegisure scan --changed --base origin/${{ github.base_ref }} --json
aegisure scan --json exits non-zero when the verdict is block, so CI can fail safely on critical findings.
A hosted dashboard beta exists for teams that want GitHub App PR comments, audit history, BYOK audit chat, and workspace views.
Learn more: https://aegisure.dev
MIT
Download counts and popularity will appear here as the collector receives them.
| Version | Published | Direct dependencies | Open version |
|---|---|---|---|
| 0.2.3 | 17 Jun 2026 | 4 |
This describes our collected data; it is not a security assessment.
Observed since 20 Oct, 20:19 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
package:73715942b67d6e4587b82d213784f0bbe0dad0c08ae0430034d6fe7073d86c2e