360 Secure SkillHub
360 安全技能中心

agent-resume

面向 OpenClaw 操作者的代理连续性 Skill:在重启、换模型或上下文清空前,通过 curl 调用 BlueColumn 的 remember 接口保存已完成与未完成工作、决策、阻塞和下一步;新会话用 recall 恢复,运行中的小决策用 note 记录。需要 BlueColumn API 密钥,数据外发至该服务。

bluecolumnconsulting-lgtm
bluecolumnconsulting-lgtm
热度 16.76
最近更新 2026-09-30 09:12:43
分类:知识依赖:第三方 API应用场景:笔记整理 / 工作流 / 任务管理

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add agent-resume
Metadata
name
agent-resume
description
Save a structured checkpoint of an agent's work before a restart or model change — completed work, unfinished tasks, decisions, blockers, next steps — and restore it into the next session. Use when an OpenClaw operator or team needs continuity across agent restarts, model swaps, or context resets. Requires a BlueColumn API key (bc_live_*).

Agent Resume — BlueColumn Skill

Paying an agent to rediscover its own work is the most expensive line item on the bill. This skill checkpoints an agent's state before any restart, model change, or context wipe, then restores it cleanly into the next session.

Setup

Get a free key at https://bluecolumn.ai/login (no card). Keys look like bc_live_*. Base URL: https://api.bluecolumn.ai.

Save a checkpoint

Before any restart, model swap, or context reset, store the state that matters.

curl -X POST https://api.bluecolumn.ai/remember \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "AGENT CHECKPOINT 2026-09-29 17:50 — Project: client onboarding automation. Completed: scraper for CRM import, field mapping docs, test run on 3 accounts (all pass). Unfinished: webhook retry logic (50% done), rate-limit handling (not started). Decisions: using Redis for retry queue, not Postgres; webhook signature = HMAC-SHA256. Blockers: staging webhook URL needs whitelisting by client IT (ticket #441). Next: finish retry logic, then ask client to whitelist staging URL.",
    "title": "agent-resume - client onboarding automation - 09-29"
  }'

Restore into the next session

On a fresh session, pull the latest checkpoint first thing.

curl -X POST https://api.bluecolumn.ai/recall \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"q": "What is the latest agent checkpoint and what was in progress, decided, blocked, and next?"}'

Log small decisions mid-run

Capture a decision the moment it happens, so it survives the next wipe.

curl -X POST https://api.bluecolumn.ai/note \
  -H "Authorization: Bearer $BLUECOLUMN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "Decision: retry queue uses Redis Streams, consumer group per tenant. Do not use Postgres SKIP LOCKED.", "tags": ["agent-resume", "decision"]}'

Workflow

  1. Checkpoint trigger — before restart, model change, context reset, or when the agent is told to "hand off".
  2. Save — completed work, unfinished work, decisions, blockers, and next steps as one structured block.
  3. Restore — new session recalls the latest checkpoint first, before doing anything else.
  4. Verify — after restore, the agent states "resuming from checkpoint: , next: " so the operator can confirm continuity.

Billing

If the API returns 402 or 429, read topup_url / upgrade_url from the error body and surface that link to the user — do not retry.

Docs

API reference: https://bluecolumn.ai/docs — fields are text, q, tags (not content/query/note).