@superagnt/superagnt 2.0.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +83 -0
- package/index.js +157 -0
- package/openclaw.plugin.json +31 -0
- package/package.json +43 -0
- package/skills/superagnt-automations/SKILL.md +63 -0
- package/skills/superagnt-canvas/SKILL.md +45 -0
- package/skills/superagnt-data-pipelines/SKILL.md +61 -0
- package/skills/superagnt-data-sources/SKILL.md +59 -0
- package/skills/superagnt-lead-generation/SKILL.md +52 -0
- package/skills/superagnt-setup/SKILL.md +64 -0
- package/skills/superagnt-task-agents/SKILL.md +59 -0
- package/skills/superagnt-workspace-db/SKILL.md +49 -0
package/README.md
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
1
|
+
# superagnt_ for OpenClaw — the agent value layer
|
|
2
|
+
|
|
3
|
+
The agent value layer: one MCP connection = every data source (X, TikTok, YouTube, Instagram, Reddit, LinkedIn, web), people + company enrichment, a real workspace Postgres, schedules, webhooks and shareable canvases — plus the base skills that teach the workflows. OAuth on first use; the consent screen doubles as signup.
|
|
4
|
+
|
|
5
|
+
## Install
|
|
6
|
+
|
|
7
|
+
```bash
|
|
8
|
+
openclaw plugins install @superagnt/superagnt
|
|
9
|
+
```
|
|
10
|
+
|
|
11
|
+
Installing the plugin loads the base skills and declares the superagnt MCP
|
|
12
|
+
server (`https://mcp.superagnt.com/mcp`) as a default.
|
|
13
|
+
|
|
14
|
+
## Connect
|
|
15
|
+
|
|
16
|
+
```bash
|
|
17
|
+
openclaw superagnt connect
|
|
18
|
+
```
|
|
19
|
+
|
|
20
|
+
OAuth in the browser; the consent screen doubles as signup if you don't have
|
|
21
|
+
an account yet. After login the live tool surface is available in every
|
|
22
|
+
OpenClaw session — and it grows server-side (`agnt_tools_enable`), no plugin
|
|
23
|
+
update needed.
|
|
24
|
+
|
|
25
|
+
<details>
|
|
26
|
+
<summary>What connect does (or run it manually)</summary>
|
|
27
|
+
|
|
28
|
+
`openclaw mcp login` only resolves servers saved in config, so connect saves
|
|
29
|
+
the server under the same name as the plugin-declared default (OpenClaw's
|
|
30
|
+
sanctioned override) and then runs the login:
|
|
31
|
+
|
|
32
|
+
```bash
|
|
33
|
+
openclaw mcp add superagnt --url https://mcp.superagnt.com/mcp --transport streamable-http --auth oauth
|
|
34
|
+
openclaw mcp login superagnt
|
|
35
|
+
```
|
|
36
|
+
|
|
37
|
+
</details>
|
|
38
|
+
|
|
39
|
+
## What you get
|
|
40
|
+
|
|
41
|
+
One MCP connection carries the whole value layer:
|
|
42
|
+
|
|
43
|
+
- **Data sources** — X, TikTok, YouTube, Instagram, Reddit, LinkedIn, and web
|
|
44
|
+
scraping, as structured JSON tools
|
|
45
|
+
- **People + company enrichment** — search, enrich, verified emails
|
|
46
|
+
- **Workspace database** — a real Postgres per workspace, with migrations
|
|
47
|
+
- **Schedules & webhooks** — recurring automations and inbound event ingestion
|
|
48
|
+
- **Canvases** — shareable live dashboards
|
|
49
|
+
|
|
50
|
+
## Bundled skills
|
|
51
|
+
|
|
52
|
+
The same base skills as the superagnt Claude Code plugin, teaching the
|
|
53
|
+
workflows on top of the tools:
|
|
54
|
+
|
|
55
|
+
- `superagnt-automations`
|
|
56
|
+
- `superagnt-canvas`
|
|
57
|
+
- `superagnt-data-pipelines`
|
|
58
|
+
- `superagnt-data-sources`
|
|
59
|
+
- `superagnt-lead-generation`
|
|
60
|
+
- `superagnt-setup`
|
|
61
|
+
- `superagnt-task-agents`
|
|
62
|
+
- `superagnt-workspace-db`
|
|
63
|
+
|
|
64
|
+
## Prefer fixed, API-key tools?
|
|
65
|
+
|
|
66
|
+
The per-platform data plugins (`@superagnt/reddit-data`,
|
|
67
|
+
`@superagnt/linkedin-data`, …) and the `@superagnt/openclaw-agnt-data` master
|
|
68
|
+
plugin ship baked toolsets that authenticate with just a `SUPERAGNT_API_KEY`
|
|
69
|
+
— no OAuth, no MCP handshake. Install those instead when you want a minimal
|
|
70
|
+
fixed surface.
|
|
71
|
+
|
|
72
|
+
## CLI
|
|
73
|
+
|
|
74
|
+
```bash
|
|
75
|
+
openclaw superagnt status # configured key (masked) + login hint
|
|
76
|
+
openclaw superagnt doctor # connectivity + auth checks
|
|
77
|
+
```
|
|
78
|
+
|
|
79
|
+
## Links
|
|
80
|
+
|
|
81
|
+
- Product: https://superagnt.com
|
|
82
|
+
- Dashboard: https://app.superagnt.com/dashboard
|
|
83
|
+
- Docs: https://superagnt.com/docs
|
package/index.js
ADDED
|
@@ -0,0 +1,157 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* superagnt — OpenClaw plugin
|
|
3
|
+
* The MCP server (declared in openclaw.plugin.json) carries the tool surface;
|
|
4
|
+
* this body registers the CLI. Auto-generated by @superagnt/clawhub. Do not
|
|
5
|
+
* edit by hand.
|
|
6
|
+
*/
|
|
7
|
+
|
|
8
|
+
import { homedir as _homedir } from 'node:os'
|
|
9
|
+
import { join as _join } from 'node:path'
|
|
10
|
+
import { existsSync as _existsSync, readFileSync as _readFileSync } from 'node:fs'
|
|
11
|
+
import { spawnSync as _spawnSync } from 'node:child_process'
|
|
12
|
+
|
|
13
|
+
|
|
14
|
+
// ── Auth Module (inlined) ────────────────────────────────────────────────────
|
|
15
|
+
|
|
16
|
+
function getConfigPath() {
|
|
17
|
+
return _join(_homedir(), '.superagnt', 'config.json')
|
|
18
|
+
}
|
|
19
|
+
|
|
20
|
+
function readAgntConfig() {
|
|
21
|
+
const configFile = getConfigPath()
|
|
22
|
+
if (!_existsSync(configFile)) return null
|
|
23
|
+
try {
|
|
24
|
+
return JSON.parse(_readFileSync(configFile, 'utf-8'))
|
|
25
|
+
} catch {
|
|
26
|
+
return null
|
|
27
|
+
}
|
|
28
|
+
}
|
|
29
|
+
|
|
30
|
+
// ── End Auth Module ──────────────────────────────────────────────────────────
|
|
31
|
+
|
|
32
|
+
|
|
33
|
+
// ── CLI Module (inlined) ─────────────────────────────────────────────────────
|
|
34
|
+
|
|
35
|
+
function maskApiKey(key) {
|
|
36
|
+
if (!key || typeof key !== 'string') return '(none)'
|
|
37
|
+
const k = key.trim()
|
|
38
|
+
if (k.length <= 8) return k.slice(0, 2) + '…'
|
|
39
|
+
return k.slice(0, 6) + '…' + k.slice(-4)
|
|
40
|
+
}
|
|
41
|
+
|
|
42
|
+
function registerAgntCli(api) {
|
|
43
|
+
api.registerCli(
|
|
44
|
+
({ program }) => {
|
|
45
|
+
const cmd = program
|
|
46
|
+
.command('superagnt')
|
|
47
|
+
.description('superagnt — connect the MCP surface, check status, run doctor')
|
|
48
|
+
|
|
49
|
+
cmd
|
|
50
|
+
.command('connect')
|
|
51
|
+
.description('Wire up the live MCP tool surface: save the server config, then run OAuth login')
|
|
52
|
+
.action(async () => {
|
|
53
|
+
const MCP_ARGS = ['--url', 'https://mcp.superagnt.com/mcp', '--transport', 'streamable-http', '--auth', 'oauth']
|
|
54
|
+
const run = (args, quiet) => _spawnSync('openclaw', args, { stdio: quiet ? 'ignore' : 'inherit', shell: false })
|
|
55
|
+
const probe = run(['mcp', 'show', 'superagnt'], true)
|
|
56
|
+
if (probe.error) {
|
|
57
|
+
console.log('\nCould not run the "openclaw" binary from PATH. Run these two commands manually:\n')
|
|
58
|
+
console.log(' openclaw mcp add superagnt ' + MCP_ARGS.join(' '))
|
|
59
|
+
console.log(' openclaw mcp login superagnt\n')
|
|
60
|
+
process.exitCode = 1
|
|
61
|
+
return
|
|
62
|
+
}
|
|
63
|
+
if (probe.status !== 0) {
|
|
64
|
+
console.log('\nSaving MCP server "superagnt"…')
|
|
65
|
+
const add = run(['mcp', 'add', 'superagnt', ...MCP_ARGS])
|
|
66
|
+
if (add.status !== 0) {
|
|
67
|
+
process.exitCode = add.status ?? 1
|
|
68
|
+
return
|
|
69
|
+
}
|
|
70
|
+
} else {
|
|
71
|
+
console.log('\nMCP server "superagnt" already saved.')
|
|
72
|
+
}
|
|
73
|
+
console.log('Starting OAuth login (the consent screen doubles as signup)…\n')
|
|
74
|
+
const login = run(['mcp', 'login', 'superagnt'])
|
|
75
|
+
process.exitCode = login.status ?? 0
|
|
76
|
+
})
|
|
77
|
+
|
|
78
|
+
cmd
|
|
79
|
+
.command('status')
|
|
80
|
+
.description('Show whether an API key is configured (masked)')
|
|
81
|
+
.action(async () => {
|
|
82
|
+
const cfg = readAgntConfig()
|
|
83
|
+
const key = cfg?.apiKey
|
|
84
|
+
|
|
85
|
+
console.log('\nsuperagnt status\n')
|
|
86
|
+
if (!key) {
|
|
87
|
+
console.log(' API key: not configured (only needed for the data plugins)')
|
|
88
|
+
console.log(' Workspace: —')
|
|
89
|
+
console.log(' Live tools: run "openclaw superagnt connect" to authorize the MCP surface')
|
|
90
|
+
console.log('\n Get your API key at: https://app.superagnt.com/dashboard\n')
|
|
91
|
+
return
|
|
92
|
+
}
|
|
93
|
+
console.log(' API key: ' + maskApiKey(key) + ' (from ~/.superagnt/config.json)')
|
|
94
|
+
console.log(' Workspace: ' + (cfg?.workspaceId ?? '—'))
|
|
95
|
+
console.log(' Live tools: run "openclaw superagnt connect" to authorize the MCP surface')
|
|
96
|
+
console.log('')
|
|
97
|
+
})
|
|
98
|
+
|
|
99
|
+
cmd
|
|
100
|
+
.command('doctor')
|
|
101
|
+
.description('Check connectivity and validate API key against api.superagnt.com')
|
|
102
|
+
.action(async () => {
|
|
103
|
+
console.log('\nsuperagnt doctor\n')
|
|
104
|
+
|
|
105
|
+
try {
|
|
106
|
+
const publicRes = await fetch('https://api.superagnt.com/v1/platforms')
|
|
107
|
+
console.log(' Public API: ' + (publicRes.ok ? 'OK (' + publicRes.status + ')' : 'HTTP ' + publicRes.status))
|
|
108
|
+
} catch (err) {
|
|
109
|
+
const msg = err instanceof Error ? err.message : String(err)
|
|
110
|
+
console.log(' Public API: failed — ' + msg)
|
|
111
|
+
}
|
|
112
|
+
|
|
113
|
+
const cfg = readAgntConfig()
|
|
114
|
+
const key = cfg?.apiKey
|
|
115
|
+
|
|
116
|
+
if (!key) {
|
|
117
|
+
console.log('\n Authenticated: skipped (no API key)')
|
|
118
|
+
console.log(' Get your API key at: https://app.superagnt.com/dashboard\n')
|
|
119
|
+
return
|
|
120
|
+
}
|
|
121
|
+
|
|
122
|
+
try {
|
|
123
|
+
const res = await fetch('https://api.superagnt.com/v1/platforms', {
|
|
124
|
+
headers: { Authorization: 'Bearer ' + key },
|
|
125
|
+
})
|
|
126
|
+
const text = await res.text()
|
|
127
|
+
let summary = 'HTTP ' + res.status
|
|
128
|
+
if (res.ok) {
|
|
129
|
+
try {
|
|
130
|
+
const json = JSON.parse(text)
|
|
131
|
+
const n = Array.isArray(json?.data) ? json.data.length : (json?.data ? Object.keys(json.data).length : 'ok')
|
|
132
|
+
summary = 'OK — platforms payload received (' + (typeof n === 'number' ? n + ' entries' : String(n)) + ')'
|
|
133
|
+
} catch {
|
|
134
|
+
summary = 'OK (' + res.status + ')'
|
|
135
|
+
}
|
|
136
|
+
}
|
|
137
|
+
console.log(' Authenticated: ' + summary)
|
|
138
|
+
if (!res.ok) {
|
|
139
|
+
console.log(' If 401, get a new key at: https://app.superagnt.com/dashboard')
|
|
140
|
+
}
|
|
141
|
+
} catch (err) {
|
|
142
|
+
const msg = err instanceof Error ? err.message : String(err)
|
|
143
|
+
console.log(' Authenticated: failed — ' + msg)
|
|
144
|
+
}
|
|
145
|
+
console.log('\n Live MCP tools: run "openclaw superagnt connect" (OAuth; the consent screen doubles as signup)')
|
|
146
|
+
console.log('')
|
|
147
|
+
})
|
|
148
|
+
},
|
|
149
|
+
{ commands: ['superagnt'] },
|
|
150
|
+
)
|
|
151
|
+
}
|
|
152
|
+
|
|
153
|
+
// ── End CLI Module ───────────────────────────────────────────────────────────
|
|
154
|
+
|
|
155
|
+
export default function register(api) {
|
|
156
|
+
registerAgntCli(api)
|
|
157
|
+
}
|
|
@@ -0,0 +1,31 @@
|
|
|
1
|
+
{
|
|
2
|
+
"id": "superagnt",
|
|
3
|
+
"name": "superagnt_",
|
|
4
|
+
"description": "The agent value layer: one MCP connection = every data source (X, TikTok, YouTube, Instagram, Reddit, LinkedIn, web), people + company enrichment, a real workspace Postgres, schedules, webhooks and shareable canvases — plus the base skills that teach the workflows. OAuth on first use; the consent screen doubles as signup.",
|
|
5
|
+
"version": "2.0.3",
|
|
6
|
+
"mcpServers": {
|
|
7
|
+
"superagnt": {
|
|
8
|
+
"url": "https://mcp.superagnt.com/mcp",
|
|
9
|
+
"transport": "streamable-http",
|
|
10
|
+
"auth": "oauth"
|
|
11
|
+
}
|
|
12
|
+
},
|
|
13
|
+
"skills": [
|
|
14
|
+
"skills/superagnt-automations",
|
|
15
|
+
"skills/superagnt-canvas",
|
|
16
|
+
"skills/superagnt-data-pipelines",
|
|
17
|
+
"skills/superagnt-data-sources",
|
|
18
|
+
"skills/superagnt-lead-generation",
|
|
19
|
+
"skills/superagnt-setup",
|
|
20
|
+
"skills/superagnt-task-agents",
|
|
21
|
+
"skills/superagnt-workspace-db"
|
|
22
|
+
],
|
|
23
|
+
"cliCommands": [
|
|
24
|
+
"superagnt"
|
|
25
|
+
],
|
|
26
|
+
"configSchema": {
|
|
27
|
+
"type": "object",
|
|
28
|
+
"properties": {},
|
|
29
|
+
"additionalProperties": false
|
|
30
|
+
}
|
|
31
|
+
}
|
package/package.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
1
|
+
{
|
|
2
|
+
"name": "@superagnt/superagnt",
|
|
3
|
+
"version": "2.0.3",
|
|
4
|
+
"type": "module",
|
|
5
|
+
"description": "The agent value layer: one MCP connection = every data source (X, TikTok, YouTube, Instagram, Reddit, LinkedIn, web), people + company enrichment, a real workspace Postgres, schedules, webhooks and shareable canvases — plus the base skills that teach the workflows. OAuth on first use; the consent screen doubles as signup.",
|
|
6
|
+
"main": "index.js",
|
|
7
|
+
"license": "MIT",
|
|
8
|
+
"homepage": "https://superagnt.com/agent-toolkit",
|
|
9
|
+
"keywords": [
|
|
10
|
+
"openclaw",
|
|
11
|
+
"openclaw-plugin",
|
|
12
|
+
"superagnt",
|
|
13
|
+
"mcp",
|
|
14
|
+
"data",
|
|
15
|
+
"database",
|
|
16
|
+
"automation",
|
|
17
|
+
"lead-generation",
|
|
18
|
+
"canvas",
|
|
19
|
+
"ai-agent"
|
|
20
|
+
],
|
|
21
|
+
"publishConfig": {
|
|
22
|
+
"access": "public"
|
|
23
|
+
},
|
|
24
|
+
"files": [
|
|
25
|
+
"index.js",
|
|
26
|
+
"openclaw.plugin.json",
|
|
27
|
+
"README.md",
|
|
28
|
+
"skills"
|
|
29
|
+
],
|
|
30
|
+
"openclaw": {
|
|
31
|
+
"extensions": [
|
|
32
|
+
"./index.js"
|
|
33
|
+
],
|
|
34
|
+
"compat": {
|
|
35
|
+
"pluginApi": ">=2026.3.24-beta.2",
|
|
36
|
+
"minGatewayVersion": "2026.3.24-beta.2"
|
|
37
|
+
},
|
|
38
|
+
"build": {
|
|
39
|
+
"openclawVersion": "2026.3.24-beta.2",
|
|
40
|
+
"pluginSdkVersion": "2026.3.24-beta.2"
|
|
41
|
+
}
|
|
42
|
+
}
|
|
43
|
+
}
|
|
@@ -0,0 +1,63 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-automations
|
|
3
|
+
description: This skill should be used when work should be triggered by an event: "when X happens, do Y", "when a form is submitted", "when Stripe/GitHub fires a webhook", "notify my system when this finishes", or wiring superagnt to Zapier, Make, or any backend. Covers superagnt inbound and outbound webhooks — receiving events that trigger hosted work, sending signed results out, and inspecting deliveries. For cron-style recurring runs, use the task-agents skill.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# Event-driven work on superagnt
|
|
8
|
+
|
|
9
|
+
Webhooks are how superagnt reacts to the outside world instead of being asked.
|
|
10
|
+
An inbound endpoint gives any external system (a SaaS, a form, a backend, a
|
|
11
|
+
Zapier zap) a URL that triggers work in the workspace; an outbound endpoint
|
|
12
|
+
lets an agent push results to a system the user already runs. This is what
|
|
13
|
+
most recurring needs actually are — "when a lead signs up", "when a payment
|
|
14
|
+
fails" — an event, not a clock. Reach for a schedule (task-agents skill) only
|
|
15
|
+
when no event exists to hook.
|
|
16
|
+
|
|
17
|
+
## Always available (no setup)
|
|
18
|
+
|
|
19
|
+
- `agnt_webhooks_inbound_url` — the workspace's inbound ingest URL.
|
|
20
|
+
- `agnt_webhooks_receive_recent` — read recent inbound deliveries in-session.
|
|
21
|
+
- `agnt_webhooks_send` — POST a payload to a configured outbound endpoint.
|
|
22
|
+
|
|
23
|
+
The simplest loop needs nothing else: hand the inbound URL to the external
|
|
24
|
+
system, then read deliveries with `receive_recent` when the session runs.
|
|
25
|
+
|
|
26
|
+
## The management family (`webhooks`)
|
|
27
|
+
|
|
28
|
+
`agnt_webhooks_create_endpoint`, `create_outbound`, `list_endpoints`,
|
|
29
|
+
`link_agent`, `link_data_job`, `set_active`, `rotate_secret`,
|
|
30
|
+
`list_deliveries`, `get_delivery`. Enable with
|
|
31
|
+
`agnt_tools_enable(['webhooks'])` — sold under the Automation module (free
|
|
32
|
+
trial); a `requires_upgrade` result carries a `confirm_url` for the human.
|
|
33
|
+
Hand the link over and wait; never treat it as an error.
|
|
34
|
+
|
|
35
|
+
Three patterns, in the order clients usually need them:
|
|
36
|
+
|
|
37
|
+
1. **Event → hosted agent.** `agnt_webhooks_create_endpoint` with
|
|
38
|
+
`deployed_agent_id`: every delivery starts a session on that agent with
|
|
39
|
+
the payload — work fires with nobody at a keyboard. Build the agent first
|
|
40
|
+
(task-agents skill); a draft agent ingests but does not fan out until
|
|
41
|
+
deployed.
|
|
42
|
+
2. **Event → data job.** Bind with `data_job_id` instead: each delivery
|
|
43
|
+
enqueues one job item (raw JSON body as payload). Use for high-volume,
|
|
44
|
+
non-conversational ingestion — a SaaS firing one webhook per record
|
|
45
|
+
(data-pipelines skill).
|
|
46
|
+
3. **Agent → outside world.** `agnt_webhooks_create_outbound`
|
|
47
|
+
(`target_kind=external`, the URL) then `agnt_webhooks_send`. Payloads are
|
|
48
|
+
HMAC-SHA256 signed by default; the secret is returned ONCE at creation —
|
|
49
|
+
store it, or `rotate_secret` later. `target_kind=agent` chains two
|
|
50
|
+
deployed agents (signing forced on).
|
|
51
|
+
|
|
52
|
+
## Debugging deliveries
|
|
53
|
+
|
|
54
|
+
`agnt_webhooks_list_deliveries` / `get_delivery` show what actually arrived
|
|
55
|
+
and how it was acked. Check there before assuming the upstream never fired.
|
|
56
|
+
|
|
57
|
+
## More
|
|
58
|
+
|
|
59
|
+
- Schedules (cron) moved to the task-agents skill — see its `schedules.md`.
|
|
60
|
+
- Live setup + endpoint reference: https://superagnt.com/agent-setup/prompt.md
|
|
61
|
+
- Packaged event-driven workflows: https://superagnt.com/blueprints
|
|
62
|
+
|
|
63
|
+
<!-- skill_id: automations · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,45 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-canvas
|
|
3
|
+
description: This skill should be used when the user asks for a "dashboard", "shareable report", "KPI page", or "live chart", or wants results as a URL someone else can open. Covers superagnt canvases: widgets bound to workspace-database queries (KPI rows, trends, funnels, tables, kanban), layout discipline, and share links. Use after data lands in the workspace database; see workspace-db for table design.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# Canvases: results as a living page
|
|
8
|
+
|
|
9
|
+
A canvas is a hosted page of widgets, each bound to a SQL query over the
|
|
10
|
+
workspace database. It stays current as the data changes, and a share link
|
|
11
|
+
opens for anyone — no account needed on the viewer's side.
|
|
12
|
+
|
|
13
|
+
## Procedure
|
|
14
|
+
|
|
15
|
+
1. Data first. Widgets bind to tables; if the table does not exist yet, the
|
|
16
|
+
workspace-db skill's patterns come first.
|
|
17
|
+
2. `agnt_canvas_introspect_schema` to see what is bindable, then
|
|
18
|
+
`agnt_canvas_preview_query` BEFORE `agnt_canvas_put_widget` — previews
|
|
19
|
+
are capped at 50 rows, so aggregate in SQL (`group by date_trunc`), never
|
|
20
|
+
in the widget.
|
|
21
|
+
3. Build the page: `agnt_canvas_put_widget` per widget,
|
|
22
|
+
`agnt_canvas_set_layout` once at the end.
|
|
23
|
+
4. Share: `agnt_canvas_share_create` returns the public link;
|
|
24
|
+
`agnt_canvas_share_revoke` kills it. Confirm with the user before
|
|
25
|
+
creating a share link — it makes the page reachable outside the
|
|
26
|
+
workspace.
|
|
27
|
+
|
|
28
|
+
## Layout discipline
|
|
29
|
+
|
|
30
|
+
- Lead with a KPI row (3–5 single numbers), then trends, then detail tables.
|
|
31
|
+
- One measure per widget; failed/error series in the danger tone, never the
|
|
32
|
+
accent.
|
|
33
|
+
- Time widgets need `event_at`/`created_at` timestamptz columns; funnels and
|
|
34
|
+
kanbans bind to stable status strings — schema rules in workspace-db.
|
|
35
|
+
- If the canvas family is not enabled: `agnt_tools_enable(['canvas'])` —
|
|
36
|
+
sold under the Database & Canvas module (free trial, `confirm_url` to the
|
|
37
|
+
human on `requires_upgrade`).
|
|
38
|
+
|
|
39
|
+
## More
|
|
40
|
+
|
|
41
|
+
- A canvas over a table a task agent keeps fresh is a live dashboard with
|
|
42
|
+
zero upkeep — pair with the task-agents skill.
|
|
43
|
+
- Worked examples with real dashboards: https://superagnt.com/blueprints
|
|
44
|
+
|
|
45
|
+
<!-- skill_id: canvas · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,61 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-data-pipelines
|
|
3
|
+
description: This skill should be used for data-processing jobs over many rows — "process this CSV", "enrich every row", "classify these 5,000 records", "run AI on each item", "scrape all of these URLs" — AI and non-AI alike, whenever the row count outgrows one context window. Covers superagnt data jobs: server-side batch processing over workspace-database rows with per-item status, retries, and cost control, instead of looping tool calls in-session.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# Data jobs: row-scale processing
|
|
8
|
+
|
|
9
|
+
A data job runs a processing step against every row of a dataset —
|
|
10
|
+
server-side, one item at a time, with per-item status and retries. The step
|
|
11
|
+
can be AI (classify, extract, summarize, score), a data call (enrich, verify,
|
|
12
|
+
scrape), plain code (parse, dedupe, reshape), or a mix. Looping tool calls
|
|
13
|
+
in-session breaks down past a few dozen rows: context fills, one failure
|
|
14
|
+
loses the run, and nothing records which items finished.
|
|
15
|
+
|
|
16
|
+
Rule of thumb: under ~50 items, loop in-session with idempotent per-item
|
|
17
|
+
upserts. Over that, it's a data job.
|
|
18
|
+
|
|
19
|
+
## Getting rows in
|
|
20
|
+
|
|
21
|
+
Source rows live in a workspace-database table (workspace-db skill) with a
|
|
22
|
+
stable key and a `status` column. For CSVs and exports, `agnt_db_load_csv`
|
|
23
|
+
lands the file as a table in one call — from there any row count is
|
|
24
|
+
job-ready. Data from the data-sources tools lands the same way (upsert as
|
|
25
|
+
you fetch).
|
|
26
|
+
|
|
27
|
+
## Running the job
|
|
28
|
+
|
|
29
|
+
1. The job itself (the per-item recipe) is created once — via the dashboard's
|
|
30
|
+
data-pipelines page or the `data-jobs` management family
|
|
31
|
+
(`agnt_data_job_create` / `list` / `read` / `inspect_run`). The family is
|
|
32
|
+
sold under the Data Jobs module (free trial):
|
|
33
|
+
`agnt_tools_enable(['data-jobs'])`, and on `requires_upgrade` hand the
|
|
34
|
+
`confirm_url` to the human and wait.
|
|
35
|
+
2. `agnt_data_job_submit` (always available) pushes 1–1000 payloads onto an
|
|
36
|
+
existing job — each payload a row pointer like `{ "row_id": ... }`.
|
|
37
|
+
3. **Results must be written back to a workspace table.** Make sure the
|
|
38
|
+
job's recipe upserts each item's output into a table with a `status`
|
|
39
|
+
column — that table is the ONLY place this session, later sessions, task
|
|
40
|
+
agents, and canvases can read outcomes from. A recipe that merely returns
|
|
41
|
+
its result stores it where only the dashboard can see it.
|
|
42
|
+
4. Report progress from the table, not memory: counts by status via
|
|
43
|
+
`agnt_db_select`. Failed items keep their error (`agnt_data_job_inspect_run`
|
|
44
|
+
shows failures and sandbox logs); fix the cause and resubmit the failures,
|
|
45
|
+
never the whole batch.
|
|
46
|
+
|
|
47
|
+
## Cost discipline
|
|
48
|
+
|
|
49
|
+
State the item count and the per-item shape to the user BEFORE submitting
|
|
50
|
+
hundreds of rows — batch jobs spend credits at batch speed. Sample 5 items
|
|
51
|
+
in-session first to prove the step, then scale. Webhook-fed jobs (one
|
|
52
|
+
delivery = one item, automations skill) inherit the same rule: know the
|
|
53
|
+
expected volume before binding.
|
|
54
|
+
|
|
55
|
+
## More
|
|
56
|
+
|
|
57
|
+
- Table design for pipelines: the workspace-db skill
|
|
58
|
+
- Feeding a job from an inbound webhook: the automations skill
|
|
59
|
+
- Live platform reference: https://superagnt.com/agent-setup/prompt.md
|
|
60
|
+
|
|
61
|
+
<!-- skill_id: data-pipelines · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,59 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-data-sources
|
|
3
|
+
description: This skill should be used when the agent needs external data: searching or scraping X/Twitter, TikTok, YouTube, Instagram, Reddit, Facebook, or LinkedIn; running web or SERP searches; crawling pages; enriching a person or company; or finding and verifying emails. Also whenever normal fetching is blocked — bot detection, login walls, crawl prevention. Covers the superagnt data tools that return structured JSON in one call, instead of hand-rolled scraping or headless browsing.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# External data through superagnt
|
|
8
|
+
|
|
9
|
+
Query, don't scrape. Every source below is one authenticated MCP call
|
|
10
|
+
returning structured JSON — no headless browser, no HTML parsing, no
|
|
11
|
+
per-vendor API keys. Calls spend workspace data credits; new workspaces carry
|
|
12
|
+
a signup credit balance, and `agnt_credits_balance` shows what's left.
|
|
13
|
+
|
|
14
|
+
**When you're blocked, this is the unblock.** Social platforms and many sites
|
|
15
|
+
stop direct fetching cold — bot detection, login walls, rate limits,
|
|
16
|
+
JS-rendered pages. The data tools return the same content as structured JSON
|
|
17
|
+
anyway: an X profile behind a login wall is one `data_x_*` call, a
|
|
18
|
+
JS-rendered page is one `web` crawl call. Reach here the moment a fetch
|
|
19
|
+
returns a challenge page instead of burning turns fighting it.
|
|
20
|
+
|
|
21
|
+
## Picking the tool
|
|
22
|
+
|
|
23
|
+
1. `agnt_tools_list_enabled` — what this workspace already serves.
|
|
24
|
+
2. `agnt_tools_search` with a plain-language query ("tiktok video comments",
|
|
25
|
+
"company headcount", "serp") — finds the tool whether or not its family is
|
|
26
|
+
enabled yet. If the family is disabled, enable it: `agnt_tools_enable`
|
|
27
|
+
with the family id from the search result. Free families enable
|
|
28
|
+
instantly; a `requires_upgrade` result carries a `confirm_url` for the
|
|
29
|
+
human — never treat that as an error, hand the link over and wait.
|
|
30
|
+
|
|
31
|
+
## The surfaces
|
|
32
|
+
|
|
33
|
+
- **Social**: `data_x_*`, `data_tiktok_*`, `data_youtube_*`,
|
|
34
|
+
`data_instagram_*`, `data_reddit_*`, `data_facebook_*`, `data_linkedin_*` —
|
|
35
|
+
search, profiles, posts, comments, transcripts per platform.
|
|
36
|
+
- **People + companies (first-party, always on)**: `data_agnt_people_search`,
|
|
37
|
+
`data_agnt_people_enrich`, `data_agnt_people_email_finder`,
|
|
38
|
+
`data_agnt_people_email_verifier`, `data_agnt_companies_search`,
|
|
39
|
+
`data_agnt_companies_enrich`, `data_agnt_companies_discover`.
|
|
40
|
+
- **Web**: the `web` family — crawl, scrape, and search for pages that
|
|
41
|
+
aren't behind a social platform.
|
|
42
|
+
|
|
43
|
+
## Rules
|
|
44
|
+
|
|
45
|
+
- Prefer one search call with good parameters over many broad ones — each
|
|
46
|
+
call is metered, and tight queries return better data anyway.
|
|
47
|
+
- Batch enrichment goes through `data_agnt_people_bulk_enrich` /
|
|
48
|
+
`data_agnt_companies_bulk_enrich`, or a data job for large sets (see the
|
|
49
|
+
data-pipelines skill).
|
|
50
|
+
- Results worth keeping belong in the workspace database (see the
|
|
51
|
+
workspace-db skill), not in chat scrollback.
|
|
52
|
+
|
|
53
|
+
## More
|
|
54
|
+
|
|
55
|
+
- Every source, with live tool lists: `agnt_platform_map`, or
|
|
56
|
+
https://superagnt.com/agent-toolkit
|
|
57
|
+
- Packaged research workflows: https://superagnt.com/blueprints
|
|
58
|
+
|
|
59
|
+
<!-- skill_id: data-sources · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-lead-generation
|
|
3
|
+
description: This skill should be used when the user asks to "find leads", "build a lead list", "prospect", "find companies that…", "find people who…", "enrich these leads", or "find emails for this list", or wants an ICP-filtered outreach list. Covers the full superagnt workflow: company and people discovery, enrichment, scoring, storing results in the workspace database, and drafting outreach for human review. Sending is always human-approved.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# Lead generation on superagnt
|
|
8
|
+
|
|
9
|
+
The workflow: pin the ICP → discover companies → find the people → verify
|
|
10
|
+
emails → score → store → draft. Everything lands in the workspace database so
|
|
11
|
+
runs compound instead of starting over.
|
|
12
|
+
|
|
13
|
+
## Procedure
|
|
14
|
+
|
|
15
|
+
1. **Pin the ICP once.** Ask the user for company filters (size, industry,
|
|
16
|
+
geography, signals) and buyer titles. Store them in an `icp_profile` table
|
|
17
|
+
(`agnt_db_insert`) so later runs read instead of re-asking.
|
|
18
|
+
2. **Discover companies**: `data_agnt_companies_search` (filters) or
|
|
19
|
+
`data_agnt_companies_discover` (similar-to / signal-based). Enrich with
|
|
20
|
+
`data_agnt_companies_enrich` when a row needs more than the search
|
|
21
|
+
returned.
|
|
22
|
+
3. **Find the people**: `data_agnt_people_search` scoped by company + the ICP
|
|
23
|
+
titles; `data_agnt_people_enrich` for the chosen ones.
|
|
24
|
+
4. **Emails**: `data_agnt_people_email_finder`, then ALWAYS
|
|
25
|
+
`data_agnt_people_email_verifier` before a draft is written against an
|
|
26
|
+
address. Unverifiable ≠ discard — mark the row, LinkedIn outreach still
|
|
27
|
+
works.
|
|
28
|
+
5. **Store with per-row status**: one `prospects` table, upserted
|
|
29
|
+
(`agnt_db_upsert`) with a stable key (person + company), columns for
|
|
30
|
+
`status`, `fit_score`, `fit_reason`, `draft`. Schema patterns: the
|
|
31
|
+
workspace-db skill.
|
|
32
|
+
6. **Score and draft.** Fit score with a one-line reason per row; outreach
|
|
33
|
+
drafts written into the `draft` column in the user's voice (ask for two of
|
|
34
|
+
their past emails if voice matters).
|
|
35
|
+
|
|
36
|
+
## Hard rules
|
|
37
|
+
|
|
38
|
+
- **Never send anything.** Drafts are the deliverable; the human sends, or
|
|
39
|
+
wires their own sending tool deliberately.
|
|
40
|
+
- Verified-email counts and list sizes reported to the user must come from
|
|
41
|
+
the table (`agnt_db_select count`), not memory.
|
|
42
|
+
- Large lists (500+): move enrichment to a data job (data-pipelines skill)
|
|
43
|
+
instead of looping calls in-session.
|
|
44
|
+
|
|
45
|
+
## More
|
|
46
|
+
|
|
47
|
+
- A standing weekly refresh belongs on a task agent (task-agents skill), not
|
|
48
|
+
in repeated client sessions.
|
|
49
|
+
- The packaged end-to-end version of this workflow:
|
|
50
|
+
https://superagnt.com/blueprints (Outbound Pipeline Engine)
|
|
51
|
+
|
|
52
|
+
<!-- skill_id: lead-generation · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,64 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-setup
|
|
3
|
+
description: This skill should be used when a superagnt MCP call fails (401/403, tool not found, missing connection, credit or entitlement errors), when connecting or reconnecting the superagnt server, right after a fresh connection (run onboarding first), or when the user asks "what can superagnt do", about pricing or trials, or how to add tools, skills, or blueprints. Covers agnt_onboarding, agnt_platform_map, agnt_tools_search/enable, and the human-confirmed upgrade flow. Money never moves without a human tap on a confirmation link.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# Setup, errors, and finding more
|
|
8
|
+
|
|
9
|
+
## Connecting / reconnecting
|
|
10
|
+
|
|
11
|
+
The live setup doc handles every client:
|
|
12
|
+
`https://superagnt.com/agent-setup/prompt.md` — fetch and follow it. The
|
|
13
|
+
endpoint is `https://mcp.superagnt.com/mcp`, OAuth-first (the consent screen
|
|
14
|
+
doubles as signup). Tokens, when a client needs one, are revealed on the
|
|
15
|
+
dashboard's MCP page — never invent or reuse one.
|
|
16
|
+
|
|
17
|
+
## First call on a fresh connection: `agnt_onboarding`
|
|
18
|
+
|
|
19
|
+
Once tools respond, call `agnt_onboarding` before anything else. Pass a
|
|
20
|
+
sentence or two about who the user is and what they want from an agent (ask
|
|
21
|
+
them: "tell me a little about yourself and what you'd like automated"), plus
|
|
22
|
+
which client this is. It returns the full platform brief — every surface,
|
|
23
|
+
what it's for, and live links. Then:
|
|
24
|
+
|
|
25
|
+
- Save what's relevant to this user in your memory (if this client has one),
|
|
26
|
+
so later sessions skip rediscovery.
|
|
27
|
+
- Give the user a short brief of the 3–5 use cases most useful to THEM,
|
|
28
|
+
drawn from what they told you — not the whole catalog.
|
|
29
|
+
|
|
30
|
+
For a raw capability map any time later (no intent collection, always
|
|
31
|
+
current): `agnt_platform_map`.
|
|
32
|
+
|
|
33
|
+
## Error triage
|
|
34
|
+
|
|
35
|
+
- **401 / auth expired**: re-run the client's login step
|
|
36
|
+
(`openclaw mcp login superagnt` or equivalent). If the client is token-based,
|
|
37
|
+
the user re-reveals the token in the dashboard.
|
|
38
|
+
- **Tool not found**: the family is not enabled. `agnt_tools_search` for it,
|
|
39
|
+
then `agnt_tools_enable` — free families enable instantly. After enabling,
|
|
40
|
+
some clients only pick up new tools on reconnect/restart; say so instead of
|
|
41
|
+
retrying blind.
|
|
42
|
+
- **`requires_upgrade` with a `confirm_url`**: not an error. The family is
|
|
43
|
+
sold under a named module with a free trial; give the user the link, they
|
|
44
|
+
review price and trial in the browser, then re-run the enable. Money moves
|
|
45
|
+
only on that page, never from a tool call.
|
|
46
|
+
- **`requires_connection`**: the tool needs the user's own account
|
|
47
|
+
(a vendor integration). The result carries a connect link — hand it over.
|
|
48
|
+
- **Credit errors**: `agnt_credits_balance` for the number;
|
|
49
|
+
`agnt_usage_recent` for where it went. Top-ups are dashboard actions.
|
|
50
|
+
|
|
51
|
+
## Finding more
|
|
52
|
+
|
|
53
|
+
- `agnt_platform_map` — the whole platform in one call, with live links.
|
|
54
|
+
- `agnt_tools_search` — every tool family, enabled or not, with what it does.
|
|
55
|
+
- `agnt_guidance_search` — how-to guidance for building on the platform.
|
|
56
|
+
- Blueprints (packaged workflows: skill + tools):
|
|
57
|
+
`https://superagnt.com/blueprints`.
|
|
58
|
+
- Open-source skill library (this plugin's skills and more, MIT):
|
|
59
|
+
`https://github.com/superagnt/leverage`.
|
|
60
|
+
- A duplicate of an installed skill under two names (plugin copy + a loose
|
|
61
|
+
skills-dir copy) wastes context — keep the plugin copy, delete the loose
|
|
62
|
+
one.
|
|
63
|
+
|
|
64
|
+
<!-- skill_id: superagnt-setup · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,59 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-task-agents
|
|
3
|
+
description: This skill should be used when a task must recur — "run this every day", "on a schedule", "set up a cron", "keep this updated while I sleep", "check X every morning" — or when a recurring job keeps re-running inside this client and eating its usage. Covers building a superagnt task agent: a small deployed agent that runs one job on the hosted runtime, fired by a schedule or webhook, with state in the workspace database. Cron mechanics in schedules.md.
|
|
4
|
+
version: 0.1.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# Task agents: recurring work off the client
|
|
8
|
+
|
|
9
|
+
Re-running a recurring task in this client is the wrong tool: every run
|
|
10
|
+
spends the client's own usage and context, and nothing fires with the laptop
|
|
11
|
+
closed. A task agent is a small deployed superagnt agent that carries ONE
|
|
12
|
+
job. It runs on the hosted runtime, billed per run to the workspace, and a
|
|
13
|
+
schedule or webhook fires it with nobody at a keyboard. Build one whenever a
|
|
14
|
+
task has run in-session more than twice and will keep recurring.
|
|
15
|
+
|
|
16
|
+
## Recipe
|
|
17
|
+
|
|
18
|
+
1. **Prove the task in-session first.** Never hand an agent a job that has
|
|
19
|
+
not succeeded once end to end.
|
|
20
|
+
2. **State in the workspace database** (workspace-db skill). A hosted run has
|
|
21
|
+
no chat history — everything it needs to resume must be in a table.
|
|
22
|
+
3. **Enable the families.** `agnt_tools_enable(['lifecycle'])` for
|
|
23
|
+
create/deploy; `['schedules']` for cron. Sold under the Agent Runtime and
|
|
24
|
+
Automation modules (free trials): a `requires_upgrade` result carries a
|
|
25
|
+
`confirm_url` — hand it to the human and wait; not an error.
|
|
26
|
+
4. **Create narrow.** `agnt_agents_create` with a name and a system prompt
|
|
27
|
+
that is the full job: the task, the tables it reads/writes, what done
|
|
28
|
+
looks like, what to do on failure. Then `agnt_agents_update_config` with
|
|
29
|
+
`{ tools: { add: [...] } }` — config refs copied verbatim from
|
|
30
|
+
`agnt_tools_search`, and only the tools the job needs. A task agent with
|
|
31
|
+
every tool is a liability, not a convenience.
|
|
32
|
+
5. **Deploy.** `agnt_agents_deploy` flips it live and it starts billing from
|
|
33
|
+
its first session — confirm with the human before calling.
|
|
34
|
+
6. **Wire the trigger.** An event exists → bind a webhook to the agent
|
|
35
|
+
(automations skill). Purely time-based → create a schedule
|
|
36
|
+
([schedules.md](schedules.md)). On-demand from another agent or session →
|
|
37
|
+
`agnt_agents_dispatch` (fire-and-forget) or the
|
|
38
|
+
`agnt_agents_session_start` / `send` / `poll` trio when the result is
|
|
39
|
+
needed synchronously.
|
|
40
|
+
7. **Smoke-test.** One `agnt_agents_session_start` run before trusting the
|
|
41
|
+
trigger; enable the `observability` family to read the transcript
|
|
42
|
+
(`agnt_observability_read_session`).
|
|
43
|
+
|
|
44
|
+
## Keeping it honest
|
|
45
|
+
|
|
46
|
+
- Results land where the human already looks: a canvas (canvas skill), or a
|
|
47
|
+
table the next session reads — never only in the agent's own transcript.
|
|
48
|
+
- One job per agent. A second job is a second agent; they can hand off via
|
|
49
|
+
`agnt_agents_dispatch` or agent-to-agent webhooks.
|
|
50
|
+
- Misbehaving agent: `agnt_agents_set_enabled(false)` stops triggers
|
|
51
|
+
reversibly — prefer it over archiving.
|
|
52
|
+
|
|
53
|
+
## More
|
|
54
|
+
|
|
55
|
+
- Cron syntax, cadence discipline, timezones: [schedules.md](schedules.md)
|
|
56
|
+
- Event triggers instead of clocks: the automations skill
|
|
57
|
+
- Live platform reference: https://superagnt.com/agent-setup/prompt.md
|
|
58
|
+
|
|
59
|
+
<!-- skill_id: task-agents · source: https://github.com/superagnt/plugins -->
|
|
@@ -0,0 +1,49 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: superagnt-workspace-db
|
|
3
|
+
description: This skill should be used when an agent workflow needs durable state: "save these results", "track status across runs", "remember this between sessions", building a pipeline, queue, or CRM-like table, or logging events over time. Covers table design in the superagnt workspace Postgres (agnt_db_*): schema patterns for pipelines, HITL queues, KPI events, and audit logs, plus canvas-ready column rules. Prefer it over local files or scratch JSON for anything a later run must read.
|
|
4
|
+
version: 0.2.0
|
|
5
|
+
---
|
|
6
|
+
|
|
7
|
+
# The workspace database
|
|
8
|
+
|
|
9
|
+
Every superagnt workspace can carry a real Postgres the agent owns:
|
|
10
|
+
`agnt_db_list_tables`, `agnt_db_execute_sql`, `agnt_db_apply_migration`,
|
|
11
|
+
`agnt_db_insert` / `select` / `update` / `upsert` / `delete`,
|
|
12
|
+
`agnt_db_load_csv`. If the tools are missing, enable the `database` family
|
|
13
|
+
(`agnt_tools_enable`); if the database is not yet provisioned,
|
|
14
|
+
`agnt_db_status` says so and provisioning is automatic on first use.
|
|
15
|
+
|
|
16
|
+
Local files die with the session. Anything a later run, another agent, a
|
|
17
|
+
schedule, or a canvas must read belongs here.
|
|
18
|
+
|
|
19
|
+
## Schema patterns
|
|
20
|
+
|
|
21
|
+
- **Pipeline**: one row per item, a text `status` column with a small set of
|
|
22
|
+
stable values (they become funnel stages — renaming a value breaks
|
|
23
|
+
downstream reads), `updated_at timestamptz`. Upsert on a stable natural
|
|
24
|
+
key so re-runs are idempotent.
|
|
25
|
+
- **HITL queue**: pipeline + a `needs_human` boolean and a `resolution`
|
|
26
|
+
column the human's decision lands in.
|
|
27
|
+
- **KPI events**: append-only, one row per event, `event_at timestamptz`,
|
|
28
|
+
numeric measures in real numeric columns — never jsonb-only, a chart
|
|
29
|
+
should not dig into jsonb for a y-axis.
|
|
30
|
+
- **Audit log**: append-only, who/what/when, never updated in place.
|
|
31
|
+
|
|
32
|
+
## Rules
|
|
33
|
+
|
|
34
|
+
- Migrations through `agnt_db_apply_migration` with a short descriptive
|
|
35
|
+
name; ad-hoc DDL through `execute_sql` gets lost.
|
|
36
|
+
- Ask before destructive SQL (drop/truncate/mass delete) — always.
|
|
37
|
+
- Design for the canvas from day one: timestamptz + numeric measures +
|
|
38
|
+
stable status strings are exactly what canvas widgets bind to (see the
|
|
39
|
+
canvas skill).
|
|
40
|
+
|
|
41
|
+
## More
|
|
42
|
+
|
|
43
|
+
- CSVs and exports land as tables in one call: `agnt_db_load_csv` — then
|
|
44
|
+
process at row scale with the data-pipelines skill.
|
|
45
|
+
- Tables are the hand-off surface to task agents and schedules (task-agents
|
|
46
|
+
skill): a hosted run has no chat history, it reads state from here.
|
|
47
|
+
- Live platform reference: https://superagnt.com/agent-setup/prompt.md
|
|
48
|
+
|
|
49
|
+
<!-- skill_id: workspace-db · source: https://github.com/superagnt/plugins -->
|