v0.2.0 — Pluggable Agentic SDLC component

What's new in v0.2.0

agent-ready graduates from a local CLI + CI gate to a drop-in pluggable component for any Agentic SDLC workflow. Five new integration surfaces ship in this release.

---

SDK barrel export (#31)

import { lintTicket, loadTicketFromFile, renderMarkdown } from "@agentlane/agent-ready";
import type { LintOutput, Ticket } from "@agentlane/agent-ready/types";

package.json now has a full exports map with subpaths: ., ./adapters, ./render, ./types. See docs/sdk.md.

---

MCP server (#32)

Wire agent-ready into Claude Desktop, Cursor, or any MCP-compatible agent:

{
  "mcpServers": {
    "agent-ready": {
      "command": "npx",
      "args": ["-y", "--package=@agentlane/agent-ready", "agent-ready-mcp"]
    }
  }
}

The agent_ready_check tool accepts target, adapter, rules, and format. Returns the full LintOutput. See docs/mcp.md.

---

Telemetry sinks (#33)

Emit LintOutput to webhook, JSONL file, or OTel collector after every lint run:

output:
  sinks:
    - type: webhook
      url: https://my-collector.example.com/agent-ready
    - type: jsonl
      path: .agent-ready/runs.jsonl
    - type: otel
      endpoint: http://localhost:4318/v1/traces

Failures are fail-soft — they log a warning but never affect the exit code. See docs/telemetry.md.

---

OPA policy bridge (#34)

Delegate rule decisions to Rego policies — remote OPA server or embedded opa eval CLI:

rules:
  enforce-pii-handling:
    type: opa
    policy: policies/pii.rego
    query: data.pii.allow
    severity: error

See docs/opa.md and example policies in examples/policies/.

---

Feedback loop (#35)

Close the loop between ticket grading and agent outcomes:

RUN_ID=$(agent-ready check PROJ-123 --format json | jq -r '.run_id')
# ... agent runs ...
agent-ready feedback record --ticket-id PROJ-123 --run-id "$RUN_ID" --outcome success
agent-ready feedback report --runs .agent-ready/runs.jsonl

The report command produces a per-rule predictive value table showing which rules reliably predict agent success. LintOutput schema bumped to 1.2 to include run_id. See docs/feedback.md.

---

Bug fixes & hardening

• --duration-min now validates for a finite number (was silently storing null on non-numeric input)
• --outcome enum validated at parse time with a clear error message
• Malformed JSONL lines in the feedback ledger and runs file emit a stderr warning and are skipped (previously returned a misleading "No feedback ledger found" error)
• OPA remote path segments are now URL-encoded via encodeURIComponent
• Built-in rule loop skips entries where a user accidentally sets type: "regex" or type: "opa" on a built-in rule ID

---

Full changelog

v0.0.4...v0.2.0
206 tests passing · 0 failures

v0.0.4

What's changed

• Migrated package and repository to the agentlane org
• Package renamed from @syedshoaib/agent-ready → @agentlane/agent-ready
• OIDC trusted publisher configured for npm — no token required for future releases
• @syedshoaib/agent-ready deprecated on npm with pointer to new package

Upgrade

npm i @agentlane/agent-ready

For the GitHub Action, update your workflow:

- uses: agentlane/agent-ready@v0

v0.0.3 - Contributor attribution fix

Agent Ready v0.0.3

Privacy and packaging maintenance release.

Fixed

• Added contributor attribution normalization so GitHub/Marketplace contributor metadata no longer attributes the project to the old work GitHub account.
• Published npm package @syedshoaib/agent-ready@0.0.3 with provenance.

Usage

- uses: Schoaib/agent-ready@v0

npx @syedshoaib/agent-ready check <ticket.json>

v0.0.2 - Scoped npm package

Agent Ready v0.0.2

This release keeps the GitHub Action unchanged and publishes the CLI to npm under the scoped package name @syedshoaib/agent-ready because npm blocks the unscoped agent-ready package name as too similar to an existing package.

Install

npx @syedshoaib/agent-ready check <ticket.json>
npm i -g @syedshoaib/agent-ready
agent-ready check <ticket.json>

GitHub Action

- uses: Schoaib/agent-ready@v0
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    comment-on-issue: true
    fail-on-not-ready: true

v0.0.1 - Agent Ready MLP

Agent Ready v0.0.1

First public MLP for agent-ready: a Definition-of-Ready linter for GitHub Issues before AI coding agents start work.

Included

• Node CLI with agent-ready and story-lint bins
• Docker-based GitHub Action
• GitHub Issue adapter and local file adapter
• 10 built-in readiness rules
• Custom regex rule packs
• Text, Markdown, and JSON output
• JSON schemas for rule packs and machine-readable output
• CI, CodeQL, Scorecard, Dependabot, and hardened repository settings

Usage

- uses: Schoaib/agent-ready@v0
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    comment-on-issue: true
    fail-on-not-ready: true

Exact pin:

- uses: Schoaib/agent-ready@v0.0.1