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/josefslerka/study-kit/opponentgit clone --depth 1 https://github.com/josefslerka/study-kitWhat 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 | $0.00023 | $0.00601 |
| Opus 5 | $0.00012 | $0.00300 |
| Sonnet 5 | $0.00005 | $0.00120 |
| Haiku 4.5 | $0.00002 | $0.00060 |
Grade A, and why
opponent 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 2d 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.
What it actually says
Write the report a journal REFEREE would write on the study. This is NOT /review:
/review verifies the study is TRUE (integrity against raw_primary/); /opponent judges
whether it's WORTH PUBLISHING (contribution and positioning against raw_secondary/). Run ONLY
AFTER /review — truth before worth.
Split two planes hard:
A) WHAT YOU JUDGE YOURSELF (from the study text + raw_primary/ + raw_secondary/):
- Triviality: isn't the thesis a banal truth no one denies?
- Overstatement: does the conclusion claim more than the body shows? (compare the title/abstract claim with what's documented)
- Method: defensible or reductive? does the declared lens hold (no diagnosing the author)?
- Scope and register: right for the target journal (from
idea.md)? register not out of genre? - Coherence of the whole: does the argument hold as a whole? (at the level of construction, not single quotes — that's /review)
- Dialogue with scholarship: does the study engage what IS in
raw_secondary/, or ignore / conveniently bend a source there?
B) WHAT YOU JUDGE ONLY AGAINST raw_secondary/ (and otherwise NOT):
- Novelty: does the study fill a real gap, or retell what the loaded scholarship already says?
- Positioning: does it fit this scholarly debate?
KEY RULE (anti-fabrication of novelty): assert novelty and positioning STRICTLY against what is
in
raw_secondary/. Where the field is not loaded, DON'T GUESS — write explicitly: "against the loaded sources there's a gap X; whether wider scholarship has already filled it, I CAN'T JUDGE — the monograph dialogue is missing". Never claim "this is new" without support; never invent what an unloaded source would contain. When a judgment lacks a source, escalate to the human ("a source belongs here, get it"), not a guess.
When raw_secondary/ is empty: do NOT do plane B at all, just note in the report that
contribution and novelty can't be judged without loaded scholarship — and do plane A normally.
OUTPUT process/opponent-report.md: summary of the study (1 paragraph, factual) · claimed
contribution (as you read it from the text) · strengths · main objections (description + what the
author would have to add + severity: major/medium/minor) · minor points · suitability for the
target journal · verdict (accept / revise / reject) · 1 most important sentence. Log it. Propose
what /revise handles and what the human must (typically: get and place a source into
raw_secondary/).
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.
- 2d ago First seen · 41 lines · 23 tokens per session scan A 1f9f23ab7f8c
opponent is a command published in the GitHub repository josefslerka/study-kit (5 stars, last pushed 2mo ago), licensed MIT. It adds 23 tokens to every session and 601 once invoked, about $0.0001 per session on Opus 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 2026-08-31.
Other commands, from other repositories
git
Git operations with intelligent commit messages and workflow optimization.
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.
constitution
Create or update the project constitution from interactive or provided principle inputs.