Borrowing it
Nothing to install: this file belongs to comet-ml/opik-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/comet-ml/opik-mcp/main/.claude/commands/dogfood.mdgit clone --depth 1 https://github.com/comet-ml/opik-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/comet-ml/opik-mcp/dogfood)<a href="https://agentmods.dev/commands/comet-ml/opik-mcp/dogfood"><img src="https://agentmods.dev/badge/commands/comet-ml/opik-mcp/dogfood/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/comet-ml/opik-mcp/dogfood"><img src="https://agentmods.dev/badge/commands/comet-ml/opik-mcp/dogfood.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00023 | $0.00953 |
| Opus 5.5 | $0.00009 | $0.00381 |
| Sonnet 5.5 | $0.00005 | $0.00191 |
| Haiku 4.5 | $0.00002 | $0.00095 |
Grade A, and why
dogfood scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 24d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Test this branch the way a customer would use it, through a headless session that loads this branch's server next to main's, and tell me if it is ready to ship. Extra focus from me: $ARGUMENTS
If the branch doesn't touch src/, both servers run the same code and no
differences is the expected result. Say that in one line and still report
problems found on both.
1. Build the two servers
Nothing gets registered in any Claude config, so no restart is needed and no other session pays for these servers.
python3 scripts/dev/install_branch.py dogfood-prepare [--workspace <ws>]
This builds this branch and a fresh origin/main into their own venvs and
writes a private MCP config naming them opik-branch and opik-base, on the
same workspace. The config holds ${OPIK_API_KEY}, never the key.
2. Read what changed and what we already know
git diff origin/main...HEAD, the commit messages, and the ticket (OPIK- from the branch name, via Jira if available).- Every file in
.claude/dogfood/memory/(format in its README).
3. Pick the flows
Core flows, phrased the way a customer would ask:
- How is my project doing?
- What is broken in production?
- Compare these two experiments.
- Find the case where the judge disagreed.
- Show me the traces with errors from the last hour.
- Explain this trace.
Add 2 to 4 flows aimed straight at what the diff changed. The servers load
only in the headless run, so describe what to pick ("the largest recent
trace", "two experiments on one dataset") and let that session find real ids
with list. Don't invent ids.
4. Run the flows in a headless session
Write the flows from step 3 to a prompt file under the scratchpad or
/tmp, with these instructions for the headless session:
- Make the same calls with the same arguments through
opik-branchandopik-base. Answer each flow as a customer would expect, then compare: is the answer right for the data, do the UI links look right, what does it cost (the size header, branch vs base), does anything say more than the data supports or cut without saying so? - Put each difference in one bucket: intended (matches the diff or ticket), regression (worse on the branch), pre-existing (wrong on both), improvement (better, and not the point of the change).
- Skip these accepted findings: <paste each
acceptedentry's symptom, unless a file under itsrecheck_whenchanged in this branch; then ask for a re-check and a "same" or "changed" answer>. - Read only. No write operations.
- Return the table from step 5 and candidate memory entries.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 24d ago First seen · 98 lines · 23 tokens per session scan A fe440db79712
dogfood is a command published in the GitHub repository comet-ml/opik-mcp (223 stars, last pushed 2d ago), licensed Apache-2.0. It adds 23 tokens to every session and 953 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-25.
Other commands, from other repositories
generate
Generate maintainable pytest tests from a URL or mobile screen via mk-qa-master's analyzer.
multi-page
Smoke test multiple pages in parallel using alethiatellparallel.
smoke
Quick Alethia smoke test — navigate and assert the page loads.
ui-aqa-flow-test-implementation
Phase 6 Test Implementation of ui-aqa-flow.
ui-aqa-flow-code-analysis
Phase 3 Code Analysis of ui-aqa-flow.
ui-aqa-flow-test-report-analysis
Phase 7 Test Report Analysis of ui-aqa-flow.