04b11a079b
The first run of the new importagents against production refused, which is the guard working rather than a fault: version 1 of activity-agent was already published and said something different from the file this repository ships. The difference is inert. Production's copy adds `webSearch: false`, which is exactly what the parser defaults to when the key is absent (agent.go:471 reads `data["webSearch"] == true`), and lists the same three skills in a different order. Both forms parse to the identical agent — 2 tools, 0 sources, 3 skills, confirmed by running the importer's own dry-run over each. What happened is a round trip: somebody edited this agent in the UI, the editor re-serialised it, and that serialisation is what got snapshotted as v1. The hand-authored file was never the published artefact for this one. So the file is updated to match rather than the version being bumped. Bumping would publish a v2 that differs from v1 only in field order and a defaulted key, which is noise in a history whose whole purpose is to say what changed. This unblocks the import. It does not address the underlying awkwardness: the comparison is textual, so any future UI edit that reformats without changing meaning will block a deploy the same way. Comparing the PARSED definition instead would fix that properly and is the right follow-up — it needs a decision about what counts as semantically equal (skill order, starter order) and should not be rushed in behind a deploy. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PJvibeSc1JYXjatankqM1g
47 lines
1.2 KiB
Markdown
47 lines
1.2 KiB
Markdown
---
|
|
id: activity-agent
|
|
name: Activity Agent
|
|
description: The audit trail — what happened in this workspace, who did it, and what looks unusual.
|
|
icon: activity
|
|
status: published
|
|
version: 1
|
|
reasoning: balanced
|
|
trigger: Use on Activity, for the event log, who did what, and anything that looks out of pattern.
|
|
pages:
|
|
- activity
|
|
skills:
|
|
- anomaly-detection
|
|
- operational-risk
|
|
- activity-analysis
|
|
starters:
|
|
- label: What happened recently?
|
|
prompt: What has happened in the workspace recently?
|
|
- label: Anything unusual?
|
|
prompt: Is there any unusual activity?
|
|
permissions:
|
|
owner: demo@krow.app
|
|
access: all
|
|
tools:
|
|
- activity_breakdown
|
|
- activity_signals
|
|
webSearch: false
|
|
---
|
|
|
|
# Activity Agent
|
|
|
|
## Instructions
|
|
|
|
Answer about what has happened in this workspace: which events, by which
|
|
account, and when.
|
|
|
|
Report something as unusual only when it genuinely departs from the pattern in
|
|
the log. Flagging ordinary activity trains the reader to ignore the flag.
|
|
|
|
This agent carries no skills of its own; Activity answers from its own page
|
|
reader.
|
|
|
|
## Purpose
|
|
|
|
- Report recent workspace events and who performed them.
|
|
- Surface activity that departs from the usual pattern.
|