Tasks

Name: Tasks
Author: rvdbreemen

ASecurity

PR-level vigilance pass orchestrating existing engines over a committed range instead of the staging area. (1) Range resolution: merge-base vs origin/main; optionally gh pr view for title/body. (2) Enforcement pass: pipe range diff into bin/adr-judge --diff -. (3) Discovery pass: bin/adr-suggest with new --intent-file flag carrying git log subjects/bodies + PR body as clearly delimited untrusted "stated intent" data (catches "migrate to Redis" confessed in comm

10 stars
0 votes
0 copies
0 views
Added October 6, 2026

ai-agentsrustrailsgit

Security analysis

A100/100

Pro scans all 8 files and shows the line behind each finding

Scanned October 6, 2026

npx -y skills add rvdbreemen/adr-kit --skill tasks --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Tasks?

Add the live security badge to your README. It updates with every re-scan.

[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rvdbreemen-tasks/badge)](https://www.skillsdirectory.com/skills/rvdbreemen-tasks)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro

SKILL.md

---
id: TASK-13
title: '/adr-kit:review: PR-level decision discovery + enforcement skill'
status: To Do
assignee: []
created_date: '2026-06-12 20:32'
labels:
  - tier-2
  - agent-guardrails
  - teams
dependencies: []
references:
  - docs/research/2026-06-12-adr-landscape.md
  - bin/adr-suggest
  - bin/adr-judge
  - bin/adr-context
priority: medium
---

## Description

<!-- SECTION:DESCRIPTION:BEGIN -->
PR-level vigilance pass orchestrating existing engines over a committed range instead of the staging area. (1) Range resolution: merge-base vs origin/main; optionally gh pr view for title/body. (2) Enforcement pass: pipe range diff into bin/adr-judge --diff -. (3) Discovery pass: bin/adr-suggest with new --intent-file flag carrying git log subjects/bodies + PR body as clearly delimited untrusted "stated intent" data (catches "migrate to Redis" confessed in commit message but invisible in diff); response may carry multiple findings. (4) Dedupe gate via bin/adr-context per candidate (high score = already documented, or supersede hint). (5) Follow-through: draft Proposed ADRs via existing authoring flow, never auto-accept. Optional advisory CI comment action (never fails build, opt-in LLM posture per v0.17). Prompt-injection caution: PR bodies attacker-writable; reuse task-12 sentinel delimiting.
<!-- SECTION:DESCRIPTION:END -->

## Acceptance Criteria
<!-- AC:BEGIN -->
- [ ] #1 Skill resolves branch range en draait judge + suggest erover
- [ ] #2 adr-suggest --intent-file voegt commit/PR-tekst toe als gedelimiteerde untrusted data, backward compatible
- [ ] #3 Dedupe via adr-context voorkomt ADR-spam
- [ ] #4 Kandidaten worden als Proposed ADR-drafts gepresenteerd, nooit auto-accepted
<!-- AC:END -->

Files in this skill

task-13 - adr-kit-review-PR-level-decision-discovery-enforcement-skill.md1.7 KB
task-40.10 - Add-Cursor-local-VS-Code-Agent-Plugins-and-goose-OpenPlugin-adapters.md3.7 KB
task-40.11 - Add-and-certify-Qwen-Code-and-Gemini-CLI-extension-adapters.md3.2 KB
task-40.12 - Add-and-certify-OMP-and-Pi-package-adapters.md3.2 KB
task-40.14 - Prototype-remaining-contract-test-clients-without-advertising-support.md3.4 KB
task-40.8 - Add-and-certify-the-Kimi-Code-ADR-Kit-plugin.md3.3 KB
task-40.9 - Build-and-certify-the-shared-Kilo-Code-and-OpenCode-adapter-family.md3.5 KB