Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/ecro/embedeval/negativesgit clone --depth 1 https://github.com/Ecro/embedevalWhat 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 | $0.00000 | $0.03418 |
| Opus 5 | $0.00000 | $0.01709 |
| Sonnet 5 | $0.00000 | $0.00684 |
| Haiku 4.5 | $0.00000 | $0.00342 |
Grade A, and why
negatives 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 yesterday.
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 — 310 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/negatives — Interactive mutation-oracle authoring
Communication Protocol
- Be direct. No preamble, no ceremony, no encouragement.
- Show the user concrete content (code, check names, reference snippets) before asking questions.
- Ask ONE decision per AskUserQuestion call — never batch multiple unrelated decisions.
- Never guess judgment calls. If unclear, ask.
- Never skip the oracle gate. A negative that doesn't pass oracle is not landed.
Sequentially author negatives.py for every case that lacks one,
with interactive judgment gates where LLM cannot reliably decide.
Usage
/negatives # sync + resume — work the next pending case
/negatives --status # sync + print progress table and exit
/negatives --case TC_ID # sync + work a specific case
/negatives --category dma # sync + cycle pending cases in a category
/negatives --sync # ONLY reconcile progress file with cases/, print diff, exit
/negatives --skip TC_ID reason # mark TC as skipped with reason (won't revisit)
/negatives --reset TC_ID # re-open a done case
/negatives --init # nuke progress file and rebuild from scratch (destructive)
Every invocation runs auto-sync first — the progress file is always reconciled
with the current cases/ directory before any action. Newly added TCs (from
Yocto/Linux expansion, new categories, etc.) appear automatically as pending
without manual bookkeeping. Removed TCs are flagged as orphaned (not deleted —
may be a rename/typo).
Progress File
Location: plans/negatives-progress.json
Schema:
{
"version": 1,
"generated": "YYYY-MM-DD",
"cases": [
{
"case_id": "dma-002",
"category": "dma",
"priority": 1,
"status": "pending | in-progress | done | skipped | orphaned",
"started_at": "YYYY-MM-DD" or null,
"completed_at": "YYYY-MM-DD" or null,
"notes": "skip reason or oracle warnings" or null
}
]
}
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.
- yesterday First seen · 310 lines · 0 tokens per session scan A 05f729e95dd5
negatives is a command published in the GitHub repository Ecro/embedeval (11 stars, last pushed 26d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 3,418 tokens. 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 2026-08-31.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.