Skip to content

YC (S26) | Open Computer History | Continuously record your company computer work, map your workflows, help you find work worth automating, and power your agents' context

docs: define accountable AI contribution policy (#5863)

Ezra Ellette29ceb57e10b7
diff --git a/.github/ISSUE_TEMPLATE/bug_report.md b/.github/ISSUE_TEMPLATE/bug_report.md
index 8a8582570..f937cfda4 100644
--- a/.github/ISSUE_TEMPLATE/bug_report.md
+++ b/.github/ISSUE_TEMPLATE/bug_report.md
@@ -7,6 +7,14 @@ assignees: ''
---
+**AI assistance and human ownership**
+read the [AI-assisted contribution policy](https://github.com/screenpipe/screenpipe/blob/main/CONTRIBUTING.md#ai-assisted-contributions).
+
+- AI tool(s) used (`none` if not used):
+- autonomy level (`none`, `autocomplete`, `chat`, or `agent`):
+- what you personally verified:
+- [ ] I personally submitted this issue; its problem statement and rationale are my own.
+
**describe the bug**
brief description of the bug.
@@ -21,4 +29,3 @@ add any other context, screenshots, or error logs here.
- you can find logs in `~/.screenpipe` on macOS/Linux, and `%APPDATA%\.screenpipe` on Windows.
- you can get app logs by right clicking in the app background and clicking the console tab.
-
diff --git a/.github/ISSUE_TEMPLATE/documentation.md b/.github/ISSUE_TEMPLATE/documentation.md
index db4341600..ecc6a7dd3 100644
--- a/.github/ISSUE_TEMPLATE/documentation.md
+++ b/.github/ISSUE_TEMPLATE/documentation.md
@@ -7,6 +7,14 @@ assignees: ''
---
+**AI assistance and human ownership**
+read the [AI-assisted contribution policy](https://github.com/screenpipe/screenpipe/blob/main/CONTRIBUTING.md#ai-assisted-contributions).
+
+- AI tool(s) used (`none` if not used):
+- autonomy level (`none`, `autocomplete`, `chat`, or `agent`):
+- what you personally verified:
+- [ ] I personally submitted this issue; its problem statement and rationale are my own.
+
**describe the issue or improvement**
explain what's missing, unclear, or needs improvement in the docs.
@@ -14,4 +22,4 @@ explain what's missing, unclear, or needs improvement in the docs.
if you have specific suggestions, please outline them here.
**additional context**
-add any other context about the documentation issue here.
\ No newline at end of file
+add any other context about the documentation issue here.
diff --git a/.github/ISSUE_TEMPLATE/feature_request.md b/.github/ISSUE_TEMPLATE/feature_request.md
index b9644c464..93ea81766 100644
--- a/.github/ISSUE_TEMPLATE/feature_request.md
+++ b/.github/ISSUE_TEMPLATE/feature_request.md
@@ -7,6 +7,14 @@ assignees: ''
---
+**AI assistance and human ownership**
+read the [AI-assisted contribution policy](https://github.com/screenpipe/screenpipe/blob/main/CONTRIBUTING.md#ai-assisted-contributions).
+
+- AI tool(s) used (`none` if not used):
+- autonomy level (`none`, `autocomplete`, `chat`, or `agent`):
+- what you personally verified:
+- [ ] I personally submitted this issue; its problem statement and rationale are my own.
+
**describe the feature**
brief description of the feature you'd like.
@@ -18,4 +26,3 @@ any alternative solutions or features you've thought about?
**additional context**
add any other context or screenshots about the feature request here.
-
diff --git a/.github/ISSUE_TEMPLATE/question.md b/.github/ISSUE_TEMPLATE/question.md
index be84db7b1..22b8390a5 100644
--- a/.github/ISSUE_TEMPLATE/question.md
+++ b/.github/ISSUE_TEMPLATE/question.md
@@ -7,8 +7,16 @@ assignees: ''
---
+**AI assistance and human ownership**
+read the [AI-assisted contribution policy](https://github.com/screenpipe/screenpipe/blob/main/CONTRIBUTING.md#ai-assisted-contributions).
+
+- AI tool(s) used (`none` if not used):
+- autonomy level (`none`, `autocomplete`, `chat`, or `agent`):
+- what you personally verified:
+- [ ] I personally submitted this issue; its problem statement and rationale are my own.
+
**your question**
ask your question here. be as specific as possible.
**context**
-add any relevant context that might help us answer your question.
\ No newline at end of file
+add any relevant context that might help us answer your question.
diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md
index 15bcce189..35775ca91 100644
--- a/.github/pull_request_template.md
+++ b/.github/pull_request_template.md
@@ -15,17 +15,37 @@ related issue: #
> attach screenshots / recordings here — **never commit media into the repo.** drag the file into this box (works for anyone, browser only) and github hosts it. on the cli: attach it as a release asset — `gh release upload <tag> file.png` if you can write here, else `gh release create media file.png --repo <you>/screenpipe` on your fork — and paste the url.
+## AI assistance and human ownership
+
+read the [AI-assisted contribution policy](https://github.com/screenpipe/screenpipe/blob/main/CONTRIBUTING.md#ai-assisted-contributions).
+
+- AI tool(s) used (`none` if not used):
+- autonomy level (`none`, `autocomplete`, `chat`, or `agent`):
+- what I personally verified:
+
+- [ ] I personally submitted this PR, understand every material change, and can explain it without asking a model to answer maintainers for me.
+- [ ] The problem statement, rationale, and replies to maintainers are my own.
+
+## evidence type
+
+check every category that applies and provide its required evidence below.
+
+- [ ] UI or behavior change — before/after recording
+- [ ] Backend change — regression tests and relevant logs/results
+- [ ] Performance change — reproducible before/after benchmarks
+- [ ] Docs, CI, or pure refactor — relevant checks and an explanation; no video required
+
## before
-a screen recording of the app/cli before this change
+for UI or behavior changes, add a recording of the app/CLI before this change. otherwise write `not applicable` and explain why.
## after
-a screen recording of the app/cli after this change
+for UI or behavior changes, add a recording of the app/CLI after this change. otherwise write `not applicable` and provide the evidence required above.
## how to test
-add a few steps to test the pr in the most time efficient way.
+list the exact commands or manual steps you personally ran and their results.
1.
2.
@@ -40,4 +60,3 @@ If this PR adds or changes `#[tauri::command]` handlers or Rust types exported t
- [ ] `bun run typecheck`
Commands are auto-collected via the vendored `tauri-helper` crate — no manual handler list edits in `main.rs`.
-
diff --git a/.github/workflows/pr-media-check.yml b/.github/workflows/pr-media-check.yml
deleted file mode 100644
index d001767dd..000000000
--- a/.github/workflows/pr-media-check.yml
+++ /dev/null
@@ -1,140 +0,0 @@
-name: PR media check (potential AI slop)
-
-# screenpipe — AI that knows everything you've seen, said, or heard
-# https://screenpipe.com
-# flags PRs whose description has no before/after screen recording. uses
-# pull_request_target so it can label/comment on fork PRs (it only reads PR
-# metadata via the api — it never checks out untrusted PR code).
-
-on:
- pull_request_target:
- types: [opened, edited, reopened, ready_for_review, synchronize]
-
-permissions:
- contents: read
- pull-requests: write
- issues: write
-
-concurrency:
- group: pr-media-check-${{ github.event.pull_request.number }}
- cancel-in-progress: true
-
-jobs:
- check-video:
- runs-on: ubuntu-latest
- steps:
- - name: Check PR body for a before/after video
- uses: actions/github-script@v7
- with:
- script: |
- const owner = context.repo.owner;
- const repo = context.repo.repo;
- const pr = context.payload.pull_request;
- const issue_number = pr.number;
-
- const LABEL = "potential-ai-slop";
- const MARKER = "<!-- potential-ai-slop-check -->";
-
- // ---- detect a video in the PR description -------------------------
- // strip image markdown first, so github drag-dropped IMAGES
- // ![alt](https://github.com/user-attachments/assets/<uuid>)
- // don't count. videos are inserted by github as a BARE url on their
- // own line, so a bare user-attachments link == video.
- const body = pr.body || "";
- const withoutImages = body.replace(/!\[[^\]]*\]\([^)]*\)/g, "");
-
- const videoPatterns = [
- /<video[\s>]/i, // html5 embed
- /\.(mp4|mov|webm|m4v|avi|mkv|ogv|mpe?g|gif)(\?[^\s)"'<>]*)?/i, // video / recording file links
- /https?:\/\/(www\.)?(youtube\.com|youtu\.be|loom\.com|vimeo\.com|streamable\.com|vidyard\.com|wistia\.com|tella\.tv|screen\.studio)\b/i, // video hosts
- /https?:\/\/github\.com\/user-attachments\/assets\/[0-9a-f-]+/i, // github drag-drop (video; images stripped above)
- /https?:\/\/user-images\.githubusercontent\.com\/\S+\.(mp4|mov|webm|m4v|gif)/i, // legacy attachment host
- ];
-
- const hasVideo = videoPatterns.some((re) => re.test(withoutImages));
- core.info(`PR #${issue_number} hasVideo=${hasVideo}`);
-
- // ---- helpers ------------------------------------------------------
- async function ensureLabel() {
- try {
- await github.rest.issues.getLabel({ owner, repo, name: LABEL });
- } catch (e) {
- if (e.status !== 404) throw e;
- await github.rest.issues.createLabel({
- owner,
- repo,
- name: LABEL,
- color: "b60205",
- description: "no before/after video in pr — may be ai slop and closed",
- });
- }
- }
-
- async function findMarkerComment() {
- const comments = await github.paginate(github.rest.issues.listComments, {
- owner,
- repo,
- issue_number,
- per_page: 100,
- });
- return comments.find((c) => c.body && c.body.includes(MARKER));
- }
-
- // ---- act ----------------------------------------------------------
- const existing = await findMarkerComment();
-
- if (!hasVideo) {
- await ensureLabel();
- await github.rest.issues.addLabels({
- owner,
- repo,
- issue_number,
- labels: [LABEL],
- });
-
- if (!existing) {
- const commentBody = [
- MARKER,
- "👋 thanks for the pr!",
- "",
- "i couldn't find a **before/after screen recording** in the description. prs without a video showing the change in action are hard to review and may be **closed soon** as potential ai slop.",
- "",
- "to clear this:",
- "- add a short screen recording of the app/cli **before** and **after** your change to the pr description",
- "- drag the file into the description box — github hosts it (never commit media into the repo)",
- "- the `" + LABEL + "` label drops automatically once a video is detected",
- "",
- "if a video genuinely doesn't apply (docs / ci / pure refactor), reply here and a maintainer can remove the label.",
- ].join("\n");
-
- await github.rest.issues.createComment({
- owner,
- repo,
- issue_number,
- body: commentBody,
- });
- core.info(`Flagged PR #${issue_number} as ${LABEL}`);
- }
- } else {
- // video present — clean up any prior flag so the check self-heals.
- try {
- await github.rest.issues.removeLabel({
- owner,
- repo,
- issue_number,
- name: LABEL,
- });
- core.info(`Removed ${LABEL} from PR #${issue_number}`);
- } catch (e) {
- if (e.status !== 404) throw e; // 404 == label wasn't set, fine
- }
-
- if (existing) {
- await github.rest.issues.deleteComment({
- owner,
- repo,
- comment_id: existing.id,
- });
- core.info(`Deleted stale slop comment on PR #${issue_number}`);
- }
- }
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index d9224da39..bd53d6c41 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -213,6 +213,59 @@ next to the released app without taking over its identity or updater settings.
## how can i contribute?
+### ai-assisted contributions
+
+ai assistance is welcome, but every outside contribution needs an accountable
+human driver. these rules apply to issues, pull requests, review comments, and
+replies to maintainers:
+
+- disclose the AI tool used, its autonomy level (`autocomplete`, `chat`, or
+ `agent`), and exactly what you personally verified. write `none` when no AI
+ tool was used.
+- do not use an autonomous agent to submit an outside issue, pull request, or
+ comment. the accountable contributor must personally make the submission.
+- write the problem statement, rationale, and replies to maintainers in your
+ own words. AI may help construct tests and reproductions, but it must not
+ impersonate you.
+- understand and be able to explain every material change. do not ask a model
+ to answer maintainers on your behalf.
+- never claim manual testing you did not perform. include the exact commands
+ and results, plus real end-to-end evidence where applicable.
+- provide evidence appropriate to the change:
+ - UI or behavior changes: a before/after recording.
+ - backend changes: regression tests and relevant logs or results.
+ - performance changes: reproducible before/after benchmarks.
+ - docs, CI, and pure refactors: the relevant checks and an explanation; do
+ not fabricate a video.
+- do not submit unsolicited LLM-generated reviews of another contributor's
+ pull request.
+
+use `autocomplete` when AI only supplied inline completions, `chat` when you
+asked for suggestions or code and chose what to apply, and `agent` when AI
+planned or executed multi-step work or edited files.
+
+maintainers use process labels to identify the concrete next step:
+
+- `needs-reproduction`: provide a reproducible case and exact steps.
+- `needs-tests`: add the missing regression or coverage tests.
+- `needs-evidence`: provide the required recording, logs, results, or benchmark.
+- `needs-human-explanation`: explain the rationale or material changes yourself.
+- `needs-scope-approval`: get maintainer agreement before continuing the work.
+
+you are welcome to reopen the contribution, or submit a corrected replacement
+that links the original, after addressing that request.
+
+reopening or resubmitting without addressing the stated request results in a
+seven-calendar-day contribution cooldown, beginning when a maintainer posts the
+cooldown notice. during the cooldown, you may not open or reopen issues or pull
+requests, or submit review comments in this repository. the next submission
+after that cooldown that again reopens or resubmits without addressing a
+concrete remediation request results in a repository contribution
+ban.
+
+enforcement is based on accountable ownership, accepted scope, evidence, and
+the contributor's ability to explain the work.
+
### reporting bugs
this section guides you through submitting a bug report for screen pipe. following these guidelines helps maintainers and the community understand your report, reproduce the behavior, and find related reports.
diff --git a/ONBOARDING.md b/ONBOARDING.md
index 643501b3f..1c2f448c4 100644
--- a/ONBOARDING.md
+++ b/ONBOARDING.md
@@ -230,9 +230,7 @@ Now go find a small issue and do it properly.
### Other tips
-- Don't send SLOP: if doing UI work, try other great products (Notion, Granola, Wisprflow, ChatGPT, Claude) and see how they do things - take inspiration and make it fit in screenpipe brand style. Test your work, untested by human work has no value
+- AI assistance is welcome, but you own the contribution. follow the [AI-assisted contribution policy](CONTRIBUTING.md#ai-assisted-contributions), test the work personally, and report only evidence you actually produced
- Be creative: go beyond what other product do, take inspiration on great artists, painters, surprise use with amazing UIs
- Can a 80 years old grand pa use the feature you shipped? If not, try harder
- Measure performance, create evals, chaos engineer your code and cover all edge cases
-- Don't be afraid of sending slop, do it, send a PR and we'll see
-
diff --git a/README.md b/README.md
index f21686ffb..482a54d92 100644
--- a/README.md
+++ b/README.md
@@ -128,7 +128,7 @@ then ask claude `what did i see in the last 5 mins?` or `summarize today convers
## Contributing
-See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines, maintainers, and how to submit PRs. AI/vibe-coded PRs welcome!
+See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines, maintainers, and how to submit PRs.
Thanks to all contributors:
diff --git a/scripts/dev/README.md b/scripts/dev/README.md
index c3a38bfbb..b088f8a2d 100644
--- a/scripts/dev/README.md
+++ b/scripts/dev/README.md
@@ -95,12 +95,12 @@ coexist on a single box.
## Producing PR evidence
-Every PR gets asked for before/after evidence (the `potential-ai-slop` bot, and
-[`CONTRIBUTING.md`](../../CONTRIBUTING.md#pull-requests)). For app/UX changes you
-film the window. For backend / CLI / DB / log changes there's no window — so show
-the old behavior then the fixed behavior **in a terminal**. `pr-evidence` records
-both in one session and renders a single GIF (headless, no browser — an agent can
-run it end to end):
+Every PR must provide the evidence appropriate to its change, as described in
+[`CONTRIBUTING.md`](../../CONTRIBUTING.md#ai-assisted-contributions). App/UX
+changes need a before/after recording. Backend / CLI / DB / log changes normally
+need regression tests and relevant logs or results; when the behavior is easiest
+to demonstrate in a terminal, `pr-evidence` can record the old behavior and the
+fix in one session and render a single GIF (headless, no browser):
```bash
brew install asciinema agg
diff --git a/scripts/dev/pr-evidence b/scripts/dev/pr-evidence
index a79279fed..460b575d0 100755
--- a/scripts/dev/pr-evidence
+++ b/scripts/dev/pr-evidence
@@ -6,11 +6,12 @@
# pr-evidence — headless before/after evidence GIF for a PR.
# Runs a BEFORE command and an AFTER command in one terminal session, records it
# with asciinema, and renders a single GIF with agg. No browser, no GUI — for
-# engine / CLI / DB / log evidence, where there's no window to film. An agent can
-# run it start to finish.
+# engine / CLI / DB / log behavior that is easiest to demonstrate in a terminal.
#
-# This is the evidence the `potential-ai-slop` bot asks for on changes that have
-# no UI: show the old behavior, then the fixed behavior, in the terminal.
+# Backend work normally needs regression tests and relevant logs or results; this
+# helper is optional supporting evidence, not a fabricated video requirement.
+# If an agent runs it, the accountable contributor must personally verify the
+# recording before citing it in a PR.
#
# deps: brew install asciinema agg
# usage: