ai-policy-runtime

Generate better AI code with task-aware policies for Claude Code.

CLEARED_STATICagent-agnosticv0.1.5

Pipeline results

Coverage

CLEARED_STATIC

Tests run

7 / 8

Portability

Agent-agnostic (works beyond Claude)

Type

Plugins

Tests & results

What the pipeline checked

TestStateResult or reason

eligibilityranall eligibility checks passed

dedupranunique -- no near-duplicate found

clocran39086 code lines / 207 files

secretsranno secret patterns matched

egressran7 lexical match(es) — FLAG-INFO (C-01: semantic analysis deferred to v2)

injectionranno patterns matched (confidence: LOW_CONFIDENCE)

required-fieldsranrequired fields present: name, version, description

behaviouraldeferredbehavioural assessment deferred (v1): 2 JS/TS entry point(s); Python with non-stdlib imports (committed-only policy): [('ai_policy_runtime/cli.py', '__future__'), ('tests/test_policy_runtime.py', 'tools')]

Coverage

Not assessed

behavioural execution not run
egress patterns logged (FLAG-INFO): 7 lexical match(es)
injection scan: English-only patterns cannot cover cjk instruction surface

README

From the repository

![AI Policy Runtime banner](docs/assets/banner.png)

# AI Policy Runtime

Generate better AI code with task-aware policies for Codex, Claude Code, and OpenCode.

AI Policy Runtime gives your AI coding agent focused engineering rules for the
task at hand. It can guide implementation, review, refactoring, API design,
performance-sensitive work, Git workflows, CMake, Python, and post-task
refinement without making you rewrite prompts for every request.

The result is more consistent agent behavior across a workspace.

## How it works

AI Policy Runtime sits between the user's task and the coding agent. It reads
the current request, workspace signals, selected packs, and Skill DSL assets,
then compiles the relevant policy into a small task-specific rule set.

The agent does not receive the full Skill library. It receives the generated
Effective Rules for the current task, rendered as prompt text and structured
JSON. Those same rules can also be used after the agent acts for review,
refinement, and repair.

## The Basic Workflow

```text
User Task
  ↓
Task Analyzer
  ↓
Structured Context
  ↓
Policy Compilation
  - activate Skills and selected Packs
  - apply dependencies and when / unle…

← Back to Plugins