Scry sigil

Summon the logs.
Keep your flow.

One query pulls the logs you need from every corner of your cloud, cast from your terminal or a Slack thread, without ever hunting through consoles.

/scry, live · Slack & TUI
#eng-oncall·via Slack
RP
riya2:32 PM

scry — oncall session
~/oncall

What the spell can do

One incantation, every source

Datadog, CloudWatch, Loki, your gateway — one query fans out, merges, and returns in under a second.

Cast from anywhere

The TUI mid-debug, or /scry in the incident thread. Same spell, same results.

Agent-native

Agents call Scry as a tool and fetch their own evidence — inside the same wards as your team.

Never break flow

The answer lands where you asked. No consoles, no re-auth, no lost thread.

The ward

Scry never talks to your cloud directly. The ward checks scopes, holds the credentials, and logs every cast.

How a Scry cast is warded: a query from the TUI or an agent enters the Spellbook ward, which holds scopes and credentials, dispatches least-privilege calls to Datadog, CloudWatch, and Loki, and returns merged, scoped results while the audit rail records every cast.01 · cast02 · ward03 · dispatch04 · returnyou · tuiagent · slackSPELLBOOKscopes · credentialsleast privilegedatadogcloudwatchlokimerged · scopedaudit trail — every cast recorded
fig. 01 · the path of a cast: warded, dispatched, returned

For humans

Mid-incident, ask in the words you already think in. The merged trail comes back in the same breath.

~/oncall ▸ scry "why did checkout 5xx spike"
  ✦ correlating 3 sources…
  ✔ payments db pool exhausted @ 14:31

For agents

Give an agent Scry as a tool. It fetches exactly the evidence it needs — never more than its ward allows.

tool: scry
  args: { service: "payments", level: "error" }
  → 4 events · scoped to team:oncall
  → returned to agent context
Scry sigil

Stop hunting for logs. Start scrying.