Skip to content
Agentmart
Submit a tool

bob-telemetry

Use when the user asks what IBM Bob consumed or did — Bobcoins, tokens, cost per model class, per task or subagent, tools used, commands flagged as dangerous…

Skill by elpfries

IndexedMIT
Write a review

Chalk Labs

Need help deploying bob-telemetry?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/elpfries-bob-telemetry

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when the user asks what IBM Bob consumed or did — Bobcoins, tokens, cost per model class, per task or subagent, tools used, commands flagged as dangerous, files edited — read from the local database ~/.bob/db/bob.db.

Use cases: backend, productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found
Runs as
Python package

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/bob-telemetry
  • /opt/bobide/resources/app
  • /opt/ibm-bob/resources/app
  • /usr/share/bobide/resources/app
  • /usr/share/ibm-bob/resources/app
  • Your home directory (~)

What's inside

Skill name
bob-telemetry
Description
Use when the user asks what IBM Bob consumed or did — Bobcoins, tokens, cost per model class, per task or subagent, tools used, commands flagged as dangerous, files edited — read from the local database ~/.bob/db/bob.db.
License
MIT
Instructions
29 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Files (3)

  • SKILL.md
  • _bobcheck.py · scripts/
  • bob_telemetry.py · scripts/

SKILL.md (first lines)

# Bob telemetry

`python3 "<skill-dir>/scripts/bob_telemetry.py" <view> [--task PREFIX] [--since YYYY-MM-DD] [--db PATH] [--json]`

| view | answers |
|---|---|
| `summary` | total Bobcoins, split by model class |
| `tasks` | cost per task |
| `calls` | every LLM call: tokens, Bobcoins, model class |
| `tools` | tool usage and durations |
| `security` | commands flagged by the security check, or unparseable |
| `changes` | files Bob edited |
| `context` | what fills the context window |

When explaining (build 2.2.3 unless a bullet says 2.3.0; 2.3.0 = `1.126.0+bob2.3.0.20261026121409`):
- `cost` is in Bobcoins.
- A root task's cost includes its subagents, its tokens do not: never add parent and subagent rows.
- Bob stores no model name; the class is inferred from the unit price — ≈ 2.0 Bobcoins per
  million tokens = standard (`premium-ide`), ≈ 0.833 = economy (`explorer`, used by `explore`
  subagents). Any other rate means billing changed: report it.
- On bob-code 2.2.0 (checked 2026-09-28 on live traffic) and bob-code 2.2.3 (checked 2026-10-19 on
  live traffic, task 9ad66ac1), calls no longer carry token counts at all (`_meta.spend` is
  `{cost, contextTokens}` only), so the class is always `n/a` here; token columns show `n/a` rather
  than 0, and the `ctx` column (context tokens for that call) is still available. Bobcoins totals are
  unaffected. The unit price can still be measured on `contextTokens`: that is what `bob-pricing`
  does (2.0 Bobcoins per million context tokens on 2.3.0, cached tokens billed at the full rate).
- 2.3.0 (checked 2026-11-03 on live traffic, task 6cc0a0b9: three priced calls, `_meta.spend` =
  `{cost, contextTokens}`, e.g. `{"cost":0.036608,"contextTokens":18304}`; `summary`, `tasks`, `calls`,
  `tools`, `security`, `changes`, `context` all run unchanged): same shape, class still `n/a`. The Bobcoins
  per million tokens measured on 2.1.0 (≈ 2.0 standard, ≈ 0.833 economy) are 2.1.0 figures, not
  re-measured since.
- The history is short, by design (2.3.0, source: `dist/extension.js` of build
  `1.126.0+bob2.3.0.20261026121409`, `cleanupOldTasks` and `deleteTaskTrees`; first seen 2026-10-19 on a
  2.2.3 install, where the database went from 26 tasks / 147 priced calls to 17 / 17 after one restart):
  at startup of the task manager Bob deletes every root task that is not pinned and whose `updated_at`
  is older than the setting `tasks.retentionDays` (default 30; a value of 0 or less disables the purge),
  then logs `Cleaned up N tasks older than D days`. Messages and pending approvals go with them
  (`ON DELETE CASCADE`). On 2.3.0 the delete is recursive (`WITH RECURSIVE` over `parent_id`): sub-task
  trees of any depth are removed with their root; 2.2.3 removed one level of children. Sub-tasks are never
  selected on their own, they follow their root. So the database holds at best about 30 days of history and
  totals from `summary` only cover that window (`summary` prints the oldest task's creation date). To keep
  more, raise `tasks.retentionDays` or pin the task.
- Schema 013_workspace_editor_sessions (2.3.0, from the migration SQL in the bundle, confirmed in the live
  database read-only): migrations 012_task_editor_sessions and 013_workspace_editor_sessions add no table
  and no column. 012 deleted the orphan `editorSession:<task id>` rows of `key_value_store` and created two
  triggers; 013 drops them and recreates them keyed by project: `delete_workspace_editor_session` (after a
  task delete, if no other task has that `project_id`, delete `editorSession:<project_id>`) and
  `ignore_orphan_editor_session_insert` (an insert of an `editorSession:<project_id>` key with no matching
  task is silently ignored). The live `sqlite_master` has exactly those two triggers; the live
  `key_value_store` holds only `featureFlags.v1`, no `editorSession:` row, and nothing in the bundle writes
  one (the string occurs only inside the SQL), so what an editor session stores is undetermined. The
  telemetry views read none of this.

The script prints titles, commands and paths, never message bodies.

From elpfries/bob-sideshow, MIT.

More from elpfries/bob-sideshow

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Oct 19, 2026 (20 days ago)
Last release
Oct 19, 2026 (20 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from elpfries

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get bob-telemetry running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying bob-telemetry?

CallGet help
Cite this page

bob-telemetry skill. Agent Mart. https://agentmart.dev/skills/elpfries-bob-telemetry (last updated 2026-10-24)