Agents
Each agent is one Markdown file: instructions, model, skills, tools, triggers, write scope and budget (PRD 10).
Agents don't run in the demo. The runs below are samples of a week of the swarm. In your own Kano, agents wake on a schedule or an event, work in an isolated branch and send their changes to you for review.
Spend this month
$0.47
of $200 budget
Runs this month
8
1 dry runs
Success rate
86%
1 failed or cancelled
Active now
0
worker runs in another process
Feedback Collectordisabled
Pull new tickets, notes, calls and messages from the connected sources since the last cursor: call list_integrations, then sync_integration for each enabled connector that has its secrets. Syncs are idempotent and skip d
Next run
—
disabled
Success
100%
2 runs
Avg cost
$0.025
anthropic/claude-sonnet-5-5
Feedback Rationalizerenabled
Triage new feedback (category, sentiment, theme, duplicates), cluster items into feedback themes, link themes to opportunities, and write the weekly digest to feedback/digests/YYYY-Www.md.
Next run
Mon 06:00 AM
on feedback.created
Success
100%
1 runs
Avg cost
$0.090
anthropic/claude-sonnet-5-5
Metrics Analystdisabled
Refresh metric values first with sync_integration (posthog). For each active goal, OKR and initiative metric, query the source system (posthog_query, posthog_get_insight, salesforce_soql) when you need more than the sync
Next run
—
disabled
Success
0%
1 runs
Avg cost
$0.010
anthropic/claude-sonnet-5-5
Interview Synthesizerdisabled
When a new interview is added, extract insights with verbatim quotes and line ranges, JTBD and pains, link persona and opportunities, then mark the interview synthesized.
Next run
—
on artifact.created
Success
100%
1 runs
Avg cost
$0.070
anthropic/claude-sonnet-5-5
PRD Writerdisabled
Draft a PRD for the given initiative from its problem, outcome, linked evidence and the default context pack.
Next run
—
disabled
Success
—
0 runs
Avg cost
—
anthropic/claude-sonnet-5-5
Roadmap Stewarddisabled
Keep delivery honest every week. If Jira is connected, refresh roadmap items with sync_integration (jira) first. Run check_delivery and explain the findings that matter: overdue items, items that will miss their release,
Next run
—
disabled
Success
100%
1 runs
Avg cost
$0.060
anthropic/claude-sonnet-5-5
Strategy Auditordisabled
Once a month, check that the work still serves the strategy. Load the context pack, run audit_strategy and read the artifacts behind the high and medium findings. Write a report (type report, kind audit, title "Strategy
Next run
—
disabled
Success
100%
1 runs
Avg cost
$0.080
anthropic/claude-sonnet-5-5
Cost by agent this month
| Agent | Runs | Tokens | Cost |
|---|---|---|---|
| Feedback Collector | 2 | 46,000 | $0.05 |
| Feedback Rationalizer | 1 | 55,500 | $0.09 |
| Metrics Analyst | 1 | 8,320 | $0.01 |
| Interview Synthesizer | 1 | 46,800 | $0.07 |
| PRD Writer | 1 | 61,700 | $0.11 |
| Roadmap Steward | 1 | 41,300 | $0.06 |
| Strategy Auditor | 1 | 44,700 | $0.08 |