> ## Documentation Index
> Fetch the complete documentation index at: https://developers.explorium.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# The v1 → v2 upgrade skill
> An agent skill that migrates an AgentSource v1 integration to v2 for you — it audits and updates endpoints, request bodies, filters, response fields, autocomplete, async jobs, events, and webhooks.
If you use **Claude Code** or another Agent Skills–compatible coding agent, you don't have to apply the [migration guide](/v2/migration-from-v1) by hand. The **`agentsource-v1-v2-upgrade`** skill packages the whole v1 → v2 contract — endpoint moves, renamed and removed fields, filter changes, async-job semantics, events, and webhooks — as an executable playbook: the agent audits your codebase with a bundled scanner, applies the smallest complete migration, updates your tests, and reports anything that needs human review.
## Download
**⬇ Download agentsource-v1-v2-upgrade.zip** (12 KB)
The ZIP contains the skill definition, the migration-rules reference, and a residual-v1 scanner:
```plaintext theme={null}
agentsource-v1-v2-upgrade/
├── SKILL.md
├── references/ # the complete v1 → v2 migration rules
└── scripts/ # scan_v1.py — inventories leftover v1 patterns
```
## Install
Extract the ZIP into the skill directory your coding agent uses:
```shell Claude Code theme={null}
mkdir -p .claude/skills
unzip agentsource-v1-v2-upgrade.zip -d .claude/skills
```
```shell Agent Skills compatible theme={null}
mkdir -p .agents/skills
unzip agentsource-v1-v2-upgrade.zip -d .agents/skills
```
The extracted layout must be `/agentsource-v1-v2-upgrade/SKILL.md`. OpenCode discovers skills in both locations.
**Commit the extracted directory** if the whole team should have the skill, and **restart your coding agent** if the skill directory didn't exist when the current session started.
### Register it in your project instructions
Add an entry under the skills section of your agent instructions file (create a short `## Agent Skills` section if there isn't one):
```markdown CLAUDE.md theme={null}
- **agentsource-v1-v2-upgrade**: Use when migrating or auditing Explorium AgentSource API integrations from v1 to v2. Invoke `/agentsource-v1-v2-upgrade` or ask Claude to migrate the AgentSource integration.
```
```markdown AGENTS.md theme={null}
- **agentsource-v1-v2-upgrade**: Migrates and audits AgentSource API integrations from v1 to v2, including endpoint, request, response, async-job, event, and webhook changes.
```
## Run the migration
Invoke it directly in Claude Code:
```plaintext theme={null}
/agentsource-v1-v2-upgrade Migrate this project's AgentSource integration from v1 to v2.
```
Or in natural language, in any compatible agent:
```plaintext theme={null}
Use the agentsource-v1-v2-upgrade skill to migrate this project from AgentSource API v1 to v2.
Apply the changes, update tests, and report anything requiring manual review.
```
## What it does
* **Audits first** — runs the bundled `scan_v1.py` to inventory v1 endpoints, fields, filters, pagination, and event names across code, tests, fixtures, and docs.
* **Applies the full contract** — `/v1` → `/v2` paths, `/bulk_enrich` removal, the `contact_information` rename and its removed `emails` field, `has_contact_details`, consolidated autocomplete, `executive_joined_company`, webhook binding, and the rest of [What's new](/v2/whats-new).
* **Preserves behavior** — keeps sync `/enrich` semantics unless your code already orchestrates async work, and implements polling, failure, and input-limit handling whenever it chooses `/job`.
* **Verifies** — updates tests to assert the new paths, bodies, and fields, re-runs the scanner, and lists any remaining manual-review items with file and line references.
Review and commit the changes only after your project's tests and the skill's residual-v1 scanner both pass.
## Related
The same contract in human-readable form — endpoint mapping, breaking changes, deprecation timeline
Every change the skill applies, with the reasoning behind it