Interlinked

Runs the interlinked detector on a pull request's description and labels the PR when it reads as agent-written. In full and description modes it also publishes the analysis, as a PR comment by default or appended to the description inside an <!-- interlinked:start --> … <!-- interlinked:end --> block. Later runs edit that comment or block in place, and the block is stripped before analysis so it never affects the score.

name: interlinked

on:
  pull_request_target:
    types: [opened, edited, reopened]

permissions:
  contents: read
  pull-requests: write

jobs:
  analyze:
    runs-on: ubuntu-latest
    steps:
      - uses: MatteoGabriele/interlinked@v0

Use pull_request_target so the token can write to PRs from forks. The action never checks out or runs PR code, so that's safe here. Edits made with the default GITHUB_TOKEN don't trigger new workflow runs, so updating the description or comment doesn't loop.

Inputs

Input	Default	Description
github-token	${{ github.token }}	Reads the PR template, updates the description, manages the label.
allowed-users		Usernames to skip, comma-separated or as a JSON array.
trusted-author-associations	member,owner	Author associations to skip (collaborator, contributor, first_timer, first_time_contributor, member, owner), comma-separated or as a JSON array. Set to "" to analyze everyone.
mode	labels	full (description and labels), labels (labels only), description (description only) or silent (outputs and job summary only).
analysis-location	comment	Where full and description modes publish the analysis: comment (a separate PR comment, edited in place on reruns) or description (appended to the PR description). See the warning below.
label-ai	agent-written	Label to add when the verdict is ai. It's never removed automatically, so editing the description can't clear it; a maintainer removes it.
auto-close	false	Close the PR when the verdict is ai.
message-ai		Custom message added to the analysis when the verdict is ai.
message-human		Custom message added to the analysis when the verdict is human.
fail-on-ai	false	Fail the step when the verdict is ai.

These follow the same conventions as agentscan-action, so both can share one workflow configuration.

Warning

With analysis-location: description, the analysis lives in text the PR author controls. They can edit or delete it, or rewrite it to say the opposite. The label and the default comment can only be changed by maintainers, so rely on those for anything you enforce.

Outputs

verdict (ai or human), probability, confidence, and signals (JSON). The analysis is also written to the job summary.