From 9b0e36a52e64d96d3fec87e91f3f0a767100c69b Mon Sep 17 00:00:00 2001 From: Chelsea Li Date: Tue, 18 Aug 2026 05:03:10 +0000 Subject: [PATCH] Add user-confirmation gate for MEMORY_INSTRUCTIONS in managed-memory skill Step 5 defined MEMORY_INSTRUCTIONS and jumped straight to Test, so agents running the skill routinely skipped surfacing the memory prompt to the user. Add an explicit confirmation step after the prompt, mirroring the Step 1 pattern that agents already honor: a bolded imperative, a mandated AskUserQuestion, and a gate that blocks the Test step until the user signs off. Synced to all five agent-* template copies. Co-authored-by: Isaac --- .claude/skills/managed-memory/SKILL.md | 2 ++ agent-langgraph-advanced/.claude/skills/managed-memory/SKILL.md | 2 ++ agent-langgraph/.claude/skills/managed-memory/SKILL.md | 2 ++ agent-openai-advanced/.claude/skills/managed-memory/SKILL.md | 2 ++ .../.claude/skills/managed-memory/SKILL.md | 2 ++ agent-openai-agents-sdk/.claude/skills/managed-memory/SKILL.md | 2 ++ 6 files changed, 12 insertions(+) diff --git a/.claude/skills/managed-memory/SKILL.md b/.claude/skills/managed-memory/SKILL.md index aca584c1..36512e2d 100644 --- a/.claude/skills/managed-memory/SKILL.md +++ b/.claude/skills/managed-memory/SKILL.md @@ -413,6 +413,8 @@ Save only what will still matter in a future, unrelated conversation — a stabl - Briefly tell the user whenever you save, update, or delete.""" ``` +**Then confirm the instructions with the user before moving on — don't skip this.** Show them the full `MEMORY_INSTRUCTIONS` text you added and explain in a sentence or two what it controls: when the agent *recalls* memory (personalized questions) and what it *saves* (durable preferences, facts, and decisions — not passing chatter). **Ask with `AskUserQuestion`** whether to keep it as-is or change it — offer options like *"Looks good"*, *"Adjust when it saves/recalls"*, and *"I'll rewrite it"*. Apply whatever they ask for, matching the wording to the scope you chose in Step 1. **Don't proceed to Test until they've signed off.** + ## Test Run the server for API-only testing with `uv run start-app --no-ui --port 8000` — plain `start-app` also clones and builds the Next.js chat UI (slow, and unneeded for curl); `--no-ui` skips it and `--port` sets the port (match it in the curls below). diff --git a/agent-langgraph-advanced/.claude/skills/managed-memory/SKILL.md b/agent-langgraph-advanced/.claude/skills/managed-memory/SKILL.md index aca584c1..36512e2d 100644 --- a/agent-langgraph-advanced/.claude/skills/managed-memory/SKILL.md +++ b/agent-langgraph-advanced/.claude/skills/managed-memory/SKILL.md @@ -413,6 +413,8 @@ Save only what will still matter in a future, unrelated conversation — a stabl - Briefly tell the user whenever you save, update, or delete.""" ``` +**Then confirm the instructions with the user before moving on — don't skip this.** Show them the full `MEMORY_INSTRUCTIONS` text you added and explain in a sentence or two what it controls: when the agent *recalls* memory (personalized questions) and what it *saves* (durable preferences, facts, and decisions — not passing chatter). **Ask with `AskUserQuestion`** whether to keep it as-is or change it — offer options like *"Looks good"*, *"Adjust when it saves/recalls"*, and *"I'll rewrite it"*. Apply whatever they ask for, matching the wording to the scope you chose in Step 1. **Don't proceed to Test until they've signed off.** + ## Test Run the server for API-only testing with `uv run start-app --no-ui --port 8000` — plain `start-app` also clones and builds the Next.js chat UI (slow, and unneeded for curl); `--no-ui` skips it and `--port` sets the port (match it in the curls below). diff --git a/agent-langgraph/.claude/skills/managed-memory/SKILL.md b/agent-langgraph/.claude/skills/managed-memory/SKILL.md index aca584c1..36512e2d 100644 --- a/agent-langgraph/.claude/skills/managed-memory/SKILL.md +++ b/agent-langgraph/.claude/skills/managed-memory/SKILL.md @@ -413,6 +413,8 @@ Save only what will still matter in a future, unrelated conversation — a stabl - Briefly tell the user whenever you save, update, or delete.""" ``` +**Then confirm the instructions with the user before moving on — don't skip this.** Show them the full `MEMORY_INSTRUCTIONS` text you added and explain in a sentence or two what it controls: when the agent *recalls* memory (personalized questions) and what it *saves* (durable preferences, facts, and decisions — not passing chatter). **Ask with `AskUserQuestion`** whether to keep it as-is or change it — offer options like *"Looks good"*, *"Adjust when it saves/recalls"*, and *"I'll rewrite it"*. Apply whatever they ask for, matching the wording to the scope you chose in Step 1. **Don't proceed to Test until they've signed off.** + ## Test Run the server for API-only testing with `uv run start-app --no-ui --port 8000` — plain `start-app` also clones and builds the Next.js chat UI (slow, and unneeded for curl); `--no-ui` skips it and `--port` sets the port (match it in the curls below). diff --git a/agent-openai-advanced/.claude/skills/managed-memory/SKILL.md b/agent-openai-advanced/.claude/skills/managed-memory/SKILL.md index aca584c1..36512e2d 100644 --- a/agent-openai-advanced/.claude/skills/managed-memory/SKILL.md +++ b/agent-openai-advanced/.claude/skills/managed-memory/SKILL.md @@ -413,6 +413,8 @@ Save only what will still matter in a future, unrelated conversation — a stabl - Briefly tell the user whenever you save, update, or delete.""" ``` +**Then confirm the instructions with the user before moving on — don't skip this.** Show them the full `MEMORY_INSTRUCTIONS` text you added and explain in a sentence or two what it controls: when the agent *recalls* memory (personalized questions) and what it *saves* (durable preferences, facts, and decisions — not passing chatter). **Ask with `AskUserQuestion`** whether to keep it as-is or change it — offer options like *"Looks good"*, *"Adjust when it saves/recalls"*, and *"I'll rewrite it"*. Apply whatever they ask for, matching the wording to the scope you chose in Step 1. **Don't proceed to Test until they've signed off.** + ## Test Run the server for API-only testing with `uv run start-app --no-ui --port 8000` — plain `start-app` also clones and builds the Next.js chat UI (slow, and unneeded for curl); `--no-ui` skips it and `--port` sets the port (match it in the curls below). diff --git a/agent-openai-agents-sdk-multiagent/.claude/skills/managed-memory/SKILL.md b/agent-openai-agents-sdk-multiagent/.claude/skills/managed-memory/SKILL.md index aca584c1..36512e2d 100644 --- a/agent-openai-agents-sdk-multiagent/.claude/skills/managed-memory/SKILL.md +++ b/agent-openai-agents-sdk-multiagent/.claude/skills/managed-memory/SKILL.md @@ -413,6 +413,8 @@ Save only what will still matter in a future, unrelated conversation — a stabl - Briefly tell the user whenever you save, update, or delete.""" ``` +**Then confirm the instructions with the user before moving on — don't skip this.** Show them the full `MEMORY_INSTRUCTIONS` text you added and explain in a sentence or two what it controls: when the agent *recalls* memory (personalized questions) and what it *saves* (durable preferences, facts, and decisions — not passing chatter). **Ask with `AskUserQuestion`** whether to keep it as-is or change it — offer options like *"Looks good"*, *"Adjust when it saves/recalls"*, and *"I'll rewrite it"*. Apply whatever they ask for, matching the wording to the scope you chose in Step 1. **Don't proceed to Test until they've signed off.** + ## Test Run the server for API-only testing with `uv run start-app --no-ui --port 8000` — plain `start-app` also clones and builds the Next.js chat UI (slow, and unneeded for curl); `--no-ui` skips it and `--port` sets the port (match it in the curls below). diff --git a/agent-openai-agents-sdk/.claude/skills/managed-memory/SKILL.md b/agent-openai-agents-sdk/.claude/skills/managed-memory/SKILL.md index aca584c1..36512e2d 100644 --- a/agent-openai-agents-sdk/.claude/skills/managed-memory/SKILL.md +++ b/agent-openai-agents-sdk/.claude/skills/managed-memory/SKILL.md @@ -413,6 +413,8 @@ Save only what will still matter in a future, unrelated conversation — a stabl - Briefly tell the user whenever you save, update, or delete.""" ``` +**Then confirm the instructions with the user before moving on — don't skip this.** Show them the full `MEMORY_INSTRUCTIONS` text you added and explain in a sentence or two what it controls: when the agent *recalls* memory (personalized questions) and what it *saves* (durable preferences, facts, and decisions — not passing chatter). **Ask with `AskUserQuestion`** whether to keep it as-is or change it — offer options like *"Looks good"*, *"Adjust when it saves/recalls"*, and *"I'll rewrite it"*. Apply whatever they ask for, matching the wording to the scope you chose in Step 1. **Don't proceed to Test until they've signed off.** + ## Test Run the server for API-only testing with `uv run start-app --no-ui --port 8000` — plain `start-app` also clones and builds the Next.js chat UI (slow, and unneeded for curl); `--no-ui` skips it and `--port` sets the port (match it in the curls below).