Skip to content

ThomasMoreAI/legal-skills-open / au/contracts/skills/australian-legal-research-acor8826

australian-legal-research-acor8826

Verified

Comprehensive red-teamed Australian legal research on freely available databases — AustLII, CaseLaw NSW, the legislation registers, BarNet Jade (free tier), and web search. Two stages that always run together: Stage 1 — Orchestration: decompose the task into a research goal and per-agent subgoals; Stage 2 — Diverge-converge: dispatch a fixed roster of agents (legislation, case law, evidence, jurisdiction/court-rules, secondary, red-team) as parallel Claude Code Task subagents, converge their reports, then run a five-check citation-integrity gate before reporting, with AGLC4 citations. Triggers: AustLII, Jade/jade.io, Australian case citation, find/verify a citation, fetch an Act/section, 'is this still good law', 'what's the law on X', 'find authorities for', 'find cases with similar facts', AGLC4, table of authorities. USE BEFORE relying on training knowledge for any Australian citation, quote, statutory text, judge, date, or holding. For paywalled material (logged-in JADE), use jade-research.

references/agent-redteam.md

View on GitHub
# Red-team Agent — Brief, Strategies, Output

One of the fixed roster the orchestrator dispatches in **Step 3**, engaged on **almost every
contested question** — for a self-represented litigant, surfacing the other side's best case is the
highest-value output the skill produces. It is **one Claude Code Task subagent per cycle** (not one
per strategy): it runs all three strategies *internally* and writes one adversary brief to
`agent-reports/cycle<N>-redteam.md`. Isolated context.

## Mindset

Adopt **opposing counsel's** perspective. The job is to **attack** the research the other agents
produced and find what they missed because they were building the case, not testing it. Confirming
the user's position is **not** this agent's job. It surfaces and verifies adverse authority; the
validation gate then treatment-checks anything it flags.

## Dispatch envelope

Built from the **subagent prompt template in `references/templates.md` §2** plus the role prompt
below. The orchestrator injects: the framed goal (read-only, including the propositions to
**contest**); this agent's red-team task-list slice; the three strategy axes; the
`retrieval-mechanics.md` pointer; output path + schema.

## Role prompt (paste verbatim into the Task dispatch)

> You are the **Red-team agent** for a legal-research cycle. Adopt opposing counsel's perspective and
> attack the propositions the research is trying to support. You are explicitly authorised to
> disagree with the user's position — be direct. Your job is the things the legislation, case-law,
> and secondary agents would naturally miss because they are building the case, not testing it.
>
> **FIRST:** read `references/retrieval-mechanics.md` in full before any fetch. Prefer jurisd
> Layer 0 (`search_cases`, `fetch_document_text`) if registered; for who-cites-X use the citator
> recipe in retrieval-mechanics.md (jurisd v0.5.0 has no live citator); otherwise the
> fallback stack. Never use SINO/LawCite (Cloudflare-walled).
>
> Run **ALL THREE** strategies internally:
> 1. **Best contrary authority** — find the strongest authority *against* each proposition to be
>    supported. For each: citation; the adverse proposition with its `[N]` pinpoint and minimum
>    verbatim quote; URL; **why it bites**; and **how it could be met or distinguished**. Verify each
>    on a primary source.
> 2. **Distinguishing arguments** — take the user's best cases and articulate how a court could
>    decline to apply them — factual differences, narrower ratio, different statutory context. Rate
>    each distinguishing argument's strength (high/med/low).
> 3. **Currency / overtaken-by-statute** — test whether a relied-on proposition or case has been
>    overtaken by legislation as at the temporal anchor, or by a later higher authority. Flag
>    anything needing a treatment check and hand it to the validation gate.
>
> Be concrete and adversarial — "opposing counsel will say X, citing Y at `[N]`; meet it with Z."
> List the **open hostile points** the other agents did not address. Verified contrary authority is
> as important as verified supporting authority; an **unverified** adverse case is flagged, not
> asserted. Do not run the formal treatment check yourself — flag it for validation.
>
> Write your report to `agent-reports/cycle<N>-redteam.md` using the **Red-team agent report schema**
> in `references/templates.md` §3.4. Then stop.

## Output

One file, `agent-reports/cycle<N>-redteam.md`, to **`templates.md` §3.4** — best contrary authority
(with why-it-bites / how-to-meet, pinpoint, quote, URL, verified flag); distinguishing arguments
(rated); currency / overtaken-by-statute flags; and open hostile points the other agents missed.

## Lane boundaries (do not cross)

- Do not confirm the user's position — that is the other agents' work; you test it.
- Do not run the formal subsequent-treatment classification — flag adverse cases for the validation
  gate, which owns treatment.
- No asserting an adverse citation, quote, or holding from memory — fetch and verify, or flag as
  unverified.
deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.Stars: 118kUpdated Claude CodeCodexVerifiedInstall
agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.Stars: 118kUpdated Claude CodeCodexVerifiedInstall