Borrowing it
Nothing to install: this file belongs to dgenio/contextweaver. 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/dgenio/contextweaver/main/.github/prompts/add-eval.prompt.mdgit clone --depth 1 https://github.com/dgenio/contextweaverWrote 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/dgenio/contextweaver/add-eval)<a href="https://agentmods.dev/commands/dgenio/contextweaver/add-eval"><img src="https://agentmods.dev/badge/commands/dgenio/contextweaver/add-eval/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/dgenio/contextweaver/add-eval"><img src="https://agentmods.dev/badge/commands/dgenio/contextweaver/add-eval.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 | $0.00016 | $0.01107 |
| Opus 5.5 | $0.00006 | $0.00443 |
| Sonnet 5.5 | $0.00003 | $0.00221 |
| Haiku 4.5 | $0.00002 | $0.00111 |
Grade A, and why
add-eval 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 1mo 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 — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Add an Eval
Follow these steps in order. Each step has a clear success criterion.
References
AGENTS.md— module map, hard rules, strong patterns, coding stylebenchmarks/README.md— harness reference + baseline tablesbenchmarks/scorecard.md— committed public scorecardscripts/render_scorecard.py— deterministic scorecard rendererscripts/benchmark_delta.py— sticky PR-comment delta generator (#211)docs/agent-context/workflows.md— canonical command reference
When to use this prompt
Adding any of: new routing scorer, tokeniser change, scoring weight change,
context-pipeline stage, sensitivity rule, adapter that owns the
ResultEnvelope.summary writer, or anything else that can move
recall@k, dropped, dedup_removed, or prompt_tokens in the
benchmark output.
Steps
-
Identify the measurement surface. Decide which substrate the feature moves: gold-set queries (
benchmarks/routing_gold.json), scenario events (benchmarks/scenarios/*.jsonl), or the per-backend matrix dimensions (benchmarks/benchmark.py:_DEFAULT_MATRIX_BACKENDS/_SIZES). Success: surface is identified. -
Extend the right input. Add gold queries / scenario events / matrix cells. For gold queries: edit
benchmarks/routing_gold.json; ensure everyexpectedid exists inexamples/sample_catalog.json; add anamespacefield. For scenarios: drop a JSONL file inbenchmarks/scenarios/; the benchmark picks it up automatically. Success: input loads without error (python benchmarks/benchmark.py --output /tmp/test.jsonexits 0). -
Regenerate the matrix.
make benchmark-matrixwrites both the single-backendroutingsummary and the per-backend × per-sizerouting_matrix+routing_per_namespacekeys tobenchmarks/results/latest.json. Success:latest.jsoncontains the new cells. -
Regenerate the scorecard.
make scorecardwritesbenchmarks/scorecard.mdfromlatest.jsondeterministically. Success:git diff --quiet benchmarks/scorecard.mdpasses after two consecutive runs on the same seed (latency lines may differ run-to-run on the same hardware; recall / drops / dedup must be byte-identical).
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.
- 1mo ago First seen · 99 lines · 16 tokens per session scan A b3a406e7cddd
add-eval is a command published in the GitHub repository dgenio/contextweaver (9 stars, last pushed 3d ago), licensed Apache-2.0. It adds 16 tokens to every session and 1,107 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 2026-09-08.
Other commands, from other repositories
implement-tasks
A task implementation command that carries out the tasks in a project plan. For code, it can use TDD, a process of writing a failing test before the code that makes it pass; documents and configuration are implemented directly.
rtl
The bar this command holds work to is already stated once, in SKILL.md's non-negotiable constraint 5 and in rtl-css-engineering.md itself — not repeated here.
typography
This command is a lookup + decision, not a multi-step build — it doesn't need its own workflow file. If the request grows into "and now build the page using this type system," hand off to page or redesign.
brief
Command "brief" from TidyFactor/Styler, covering when to use and dispatch steps.
test
Run the project test suite and report results.
openrouter-live-smoke
Run an honest, credit-careful, key-safe LIVE real-boundary smoke against OpenRouter (a single model via councilinference, a council via deepseekreview preset, or the GUI proxy). Probe reachability first, gate the live call, leak-check the key, capture honest attrition, and record real-boundary-smoke in the reality…