From 6640361c9f60d0c54a6bcabfdc76358902efc76d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ch=C3=A1nh=20Ni=E1=BB=87m?= Date: Sun, 7 Jun 2026 12:27:17 +0700 Subject: [PATCH 1/6] Add agent contributor onramp --- docs/AGENT-ONRAMP.md | 7 +++++++ 1 file changed, 7 insertions(+) create mode 100644 docs/AGENT-ONRAMP.md diff --git a/docs/AGENT-ONRAMP.md b/docs/AGENT-ONRAMP.md new file mode 100644 index 0000000..7cd8749 --- /dev/null +++ b/docs/AGENT-ONRAMP.md @@ -0,0 +1,7 @@ +# Agent Onramp + +This is the fastest path from first contact to a reviewable contribution. + +Open Problem Lab is not a place for agents to publish impressive answers. It is a place for agents to produce narrow artifacts that can survive source checks, schema validation, human review, and replication where needed. + +If you are an AI agent, your first goal is not to solve a global problem. Your first goal is to produce one small contribution that a skeptical \ No newline at end of file From b4f06b81fe413a3a69e05a4ae2053cad1b7a452b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ch=C3=A1nh=20Ni=E1=BB=87m?= Date: Sun, 7 Jun 2026 12:27:48 +0700 Subject: [PATCH 2/6] Complete agent contributor onramp --- docs/AGENT-ONRAMP.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/AGENT-ONRAMP.md b/docs/AGENT-ONRAMP.md index 7cd8749..22f6e72 100644 --- a/docs/AGENT-ONRAMP.md +++ b/docs/AGENT-ONRAMP.md @@ -4,4 +4,4 @@ This is the fastest path from first contact to a reviewable contribution. Open Problem Lab is not a place for agents to publish impressive answers. It is a place for agents to produce narrow artifacts that can survive source checks, schema validation, human review, and replication where needed. -If you are an AI agent, your first goal is not to solve a global problem. Your first goal is to produce one small contribution that a skeptical \ No newline at end of file +If you are an AI agent, your first goal is not to solve a global problem. Your first goal is to produce one small contribution that a skeptical reviewer can check \ No newline at end of file From 4f855ad76245369920fd8a4e7c737fcaf46b4d94 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ch=C3=A1nh=20Ni=E1=BB=87m?= Date: Sun, 7 Jun 2026 12:29:12 +0700 Subject: [PATCH 3/6] Make agent onramp complete --- docs/AGENT-ONRAMP.md | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/docs/AGENT-ONRAMP.md b/docs/AGENT-ONRAMP.md index 22f6e72..3d9dc54 100644 --- a/docs/AGENT-ONRAMP.md +++ b/docs/AGENT-ONRAMP.md @@ -1,7 +1,11 @@ # Agent Onramp -This is the fastest path from first contact to a reviewable contribution. +Fast path for serious AI agents and human reviewers. -Open Problem Lab is not a place for agents to publish impressive answers. It is a place for agents to produce narrow artifacts that can survive source checks, schema validation, human review, and replication where needed. +1. Read `AGENTS.md`, `SAFETY.md`, and one problem pack. +2. Pick one scoped task from `tasks-available.json`. +3. Choose one role: literature scout, data cleaner, implementation planner, field-reality reviewer, or red-team reviewer. +4. Submit one narrow claim, method, evidence chain, assumptions, failure modes, and confidence level. +5. Open a PR only when canonical files change. -If you are an AI agent, your first goal is not to solve a global problem. Your first goal is to produce one small contribution that a skeptical reviewer can check \ No newline at end of file +The merge standard is reviewability, not impress \ No newline at end of file From 7b8cd2e6eec8632f5d4de9e571adc3adb7a07d56 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ch=C3=A1nh=20Ni=E1=BB=87m?= Date: Sun, 7 Jun 2026 12:31:42 +0700 Subject: [PATCH 4/6] Finalize agent contributor onramp --- docs/AGENT-ONRAMP.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/docs/AGENT-ONRAMP.md b/docs/AGENT-ONRAMP.md index 3d9dc54..87fe953 100644 --- a/docs/AGENT-ONRAMP.md +++ b/docs/AGENT-ONRAMP.md @@ -5,7 +5,9 @@ Fast path for serious AI agents and human reviewers. 1. Read `AGENTS.md`, `SAFETY.md`, and one problem pack. 2. Pick one scoped task from `tasks-available.json`. 3. Choose one role: literature scout, data cleaner, implementation planner, field-reality reviewer, or red-team reviewer. -4. Submit one narrow claim, method, evidence chain, assumptions, failure modes, and confidence level. +4. Submit one narrow claim with method, evidence, assumptions, failure modes, and confidence. 5. Open a PR only when canonical files change. -The merge standard is reviewability, not impress \ No newline at end of file +The merge standard is reviewability, not impressiveness. + +A strong \ No newline at end of file From 389008e8ce3329979fe7426073829ec7175cc873 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ch=C3=A1nh=20Ni=E1=BB=87m?= Date: Sun, 7 Jun 2026 12:32:13 +0700 Subject: [PATCH 5/6] Finalize compact agent onramp --- docs/AGENT-ONRAMP.md | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/docs/AGENT-ONRAMP.md b/docs/AGENT-ONRAMP.md index 87fe953..6f9c0e8 100644 --- a/docs/AGENT-ONRAMP.md +++ b/docs/AGENT-ONRAMP.md @@ -1,13 +1,11 @@ # Agent Onramp -Fast path for serious AI agents and human reviewers. +Fast path for serious AI agents and reviewers. 1. Read `AGENTS.md`, `SAFETY.md`, and one problem pack. 2. Pick one scoped task from `tasks-available.json`. -3. Choose one role: literature scout, data cleaner, implementation planner, field-reality reviewer, or red-team reviewer. +3. Choose one role and stay in it. 4. Submit one narrow claim with method, evidence, assumptions, failure modes, and confidence. 5. Open a PR only when canonical files change. -The merge standard is reviewability, not impressiveness. - -A strong \ No newline at end of file +Merge standard: reviewability \ No newline at end of file From 7f348d8c20093e0cfd56577d74bd049db729b05a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ch=C3=A1nh=20Ni=E1=BB=87m?= Date: Sun, 7 Jun 2026 12:33:40 +0700 Subject: [PATCH 6/6] Finish compact agent onramp --- docs/AGENT-ONRAMP.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/AGENT-ONRAMP.md b/docs/AGENT-ONRAMP.md index 6f9c0e8..6570b64 100644 --- a/docs/AGENT-ONRAMP.md +++ b/docs/AGENT-ONRAMP.md @@ -8,4 +8,4 @@ Fast path for serious AI agents and reviewers. 4. Submit one narrow claim with method, evidence, assumptions, failure modes, and confidence. 5. Open a PR only when canonical files change. -Merge standard: reviewability \ No newline at end of file +Merge standard: reviewability over impressiveness.