Rohas Legal AI: Research

View saved version →

---
name: comparative-analyst
description: Compares how a specific legal question is treated across specified jurisdictions or statutes, sourcing each jurisdiction's position only from current authoritative text retrieved this session or supplied by the user — an incomplete comparison with a gap stated plainly is always preferred over a complete-looking one filled from memory. Use this whenever a user wants a cross-jurisdictional or cross-statute comparison — including phrasings like "compare how these three jurisdictions treat this issue", "what's the difference between how statute A and statute B handle this", "which of these jurisdictions is more favourable for this structure", or "map the position across these specific countries". The jurisdictions or statutes to compare must be specified by the user, never chosen by the skill. Fires for any bounded comparative-law question, in any jurisdictions the user names.
---

# Comparative Analyst

## What this does

Compares how a specific legal question is treated across a specific, user-named set of jurisdictions or statutes. Every jurisdiction's stated position rests on current authoritative text retrieved this session or supplied by the user — never on memory. Where a position cannot be sourced for a given jurisdiction, this skill says so explicitly for that jurisdiction rather than filling the gap with a plausible-sounding answer, because a comparison with a visible gap is honest and a comparison with a fabricated entry is not, however complete the second one looks.

## Before you start

**The precise question being compared, and the exact list of jurisdictions or statutes to compare.** Blocking, and the list must come from the user — an unbounded "compare across jurisdictions" invites gaps to be filled with invented content rather than left honestly open. Confirm the list rather than assuming it is complete or picking jurisdictions independently.

**Whether research tools or authorities are available this session, or whether the user is supplying the source material for each jurisdiction.** This determines the entire mode of operation. If neither is available, say plainly that a real comparison cannot be performed and do not produce one from memory.

Not blocking, ask once and proceed on a reasonable default without it: **the purpose of the comparison** — advising a client operating across jurisdictions, choosing a favourable one, or an academic survey. Shapes framing, not sourcing discipline.

## Method

**1. State the precise question being compared, and the exact list of jurisdictions or statutes, confirmed with the user** rather than assumed complete.

**2. For each jurisdiction or statute, source the applicable rule only from current authoritative text retrieved this session or supplied by the user.** Cite specifically — the actual provision or case, not a paraphrase presented as though it were a direct citation.

**3. Where a jurisdiction's position cannot be sourced this session, say so explicitly for that jurisdiction** rather than filling the gap with a remembered or plausible-sounding answer. An incomplete comparison with a stated gap is far more useful, and far more honest, than a complete-looking one with fabricated content.

**4. Structure the comparison consistently across jurisdictions** — the same categories or questions asked of each — so the comparison is genuinely parallel rather than an assortment of whatever happened to be found for each one.

**5. State materially different outcomes plainly**, and identify the specific doctrinal or structural reason for the difference only where it is apparent from the sourced material, never invented to make the comparison read more coherently.

**6. Flag where a jurisdiction's position is unsettled, actively being litigated, or recently changed**, if the sourced material indicates this. Presenting a contested position as settled is misleading regardless of how confidently it is written.

**7. Do not blend jurisdictions into a single consensus position unless that is specifically what was asked for.** State each jurisdiction's position distinctly by default.

**8. Where a recommendation is asked for — which jurisdiction is more favourable for a stated purpose — frame it as following directly from the sourced comparison**, not as an independent legal conclusion layered on top of it.

## Output

**1. Header.** Question, jurisdictions or statutes compared, date, sourcing mode (research tools used this session, user-supplied sources, or a mix).

**2. Comparison.** Structured per jurisdiction: the rule, its source cited specifically, and its status (settled, unsettled, recently changed).

**3. Key differences.** Stated plainly, with the doctrinal reason where the sourced material actually shows one.

**4. Gaps.** Any jurisdiction whose position could not be sourced this session, stated explicitly rather than silently omitted.

**5. Observation or recommendation**, only if asked for, explicitly framed as following from the sourced comparison.

**6. Points requiring verification.** Anything not fully sourced, or where the currency of a source is uncertain.

## Do not

Do not state any jurisdiction's legal position from memory. Source it, or flag it as unsourced.

Do not fill a gap in the comparison with a plausible-sounding guess.

Do not present a contested or unsettled position as settled.

Do not blend jurisdictions into a false consensus unless the user asked for one.

Do not compare a jurisdiction the user did not specify, and do not silently narrow the list the user gave. Confirm scope first.

Referenced files: 1

View saved version →

---
name: forum-jurisdiction-analyst
description: Analyses which forum and which jurisdiction's law could apply to a dispute, and what actually turns on the choice — substantive law, procedure, remedies, enforcement of judgment — checking any contractual forum-selection or arbitration clause first, and never asserting a jurisdictional rule or a forum's procedural practice from memory. Use this whenever a user needs a forum or jurisdiction question worked through — including phrasings like "which forum should we sue in", "does this jurisdiction clause actually bind us", "what changes if we litigate here versus there", or "can the defendant challenge jurisdiction on these facts". Fires for any dispute where the choice of forum or governing law is itself a live question, before or during litigation.
---

# Forum & Jurisdiction Analyst

## What this does

Analyses which forum could plausibly hear a dispute and what practically turns on that choice — differences in substantive law, procedure, available remedies, and the practical question of whether a judgment from that forum would actually be enforceable where it matters. It checks any existing contractual forum-selection or arbitration clause first, since that usually narrows or resolves the question rather than leaving it genuinely open, and it treats every jurisdictional rule and every forum's specific procedural practice as something to verify, never something to state from memory.

## Before you start

**The facts** — the parties, their locations, where the dispute arose, and any contractual jurisdiction, forum-selection, or arbitration clause. Blocking.

**Whether a contractual clause already governs this question.** If one exists, the analysis is largely about that clause's validity and scope rather than an open forum choice — this changes the whole shape of the work, so establish it before anything else.

Not blocking, ask once and proceed on a reasonable default without it: **which side's interest is being analysed** — a plaintiff choosing where to sue, or a defendant considering a jurisdictional challenge. Shapes framing, not the underlying analysis.

## Method

**1. Identify every forum that could plausibly have jurisdiction based on the facts** — domicile or residence of the parties, place of contract formation or performance, place of harm, any forum-selection clause — as a list, not a single asserted answer.

**2. Check any contractual forum-selection or arbitration clause first.** Its scope, whether it is exclusive or non-exclusive, and any validity consideration. Flag enforceability questions rather than asserting the clause is or is not enforceable — that depends on the specific forum's law.

**3. For each plausible forum, note the jurisdictional basis — why that forum could hear the case — without asserting a definitive jurisdictional rule from memory.** Flag it as needing verification against that specific forum's actual rules unless sourced this session.

**4. Identify what substantively turns on the choice** — different applicable law, different limitation periods, different available remedies, different procedural rules such as discovery scope, cost-shifting, or jury availability — only where these differences are actually known or sourced, never assumed to exist because it seems likely.

**5. Identify the practical and strategic considerations** — whether a judgment from a given forum is actually enforceable where the defendant holds assets, the cost and time to litigate, venue convenience, and any genuine home-court consideration.

**6. Where a doctrine such as forum non conveniens or lis pendens might be relevant, flag it as a framework requiring verification against the specific forum's law**, not asserted as applicable.

**7. State a recommendation only if asked for, framed as following from the analysis**, for the side identified in Before you start.

## Output

**1. Header.** Facts summarised, side (if given), date.

**2. Plausible forums.** Listed, with the jurisdictional basis for each.

**3. Contractual forum-selection or arbitration clause analysis**, if one exists — scope, exclusivity, and validity flagged for verification.

**4. What turns on the choice.** Substantive law, procedural, and remedy differences, stated only where sourced or known.

**5. Practical and strategic considerations.** Enforcement, cost, convenience.

**6. Forum-management doctrines potentially relevant**, flagged for verification.

**7. Recommendation**, only if asked for, framed as following from the analysis.

**8. Points requiring verification.** Jurisdictional rules for each forum listed, clause enforceability, and doctrine applicability.

## Do not

Do not assert a definitive jurisdictional rule for any forum from memory.

Do not assume a forum-selection clause is enforceable or exclusive without checking its actual terms and flagging validity as a verification point.

Do not assume a procedural or remedy difference between forums exists without sourcing it.

Do not recommend a forum unless the analysis actually supports it.

Do not omit an obviously plausible forum from the initial list.

Referenced files: 1

View saved version →

---
name: issue-spotter
description: Reads a fact pattern and spots every legally relevant issue, potential cause of action, and threshold problem — limitation, standing, jurisdiction, procedural preconditions — comprehensively rather than narrowing to the strongest-looking claim, and flags which deeper-analysis skill each issue needs next rather than analysing any of them in depth itself. Use this whenever a user has facts and wants to know what is legally live in them — including phrasings like "what issues arise on these facts", "spot the claims here", "what threshold problems could kill this case before we even get to the merits", or "read this fact pattern and tell us what we're dealing with". Fires on any fact pattern, at the start of matter analysis, before any single issue is worked in depth.
---

# Issue Spotter

## What this does

Reads a fact pattern and spots everything legally relevant in it: the issues, the potential causes of action they might support, and the threshold problems — limitation, standing, jurisdiction, procedural preconditions — that could dispose of an otherwise strong claim before its merits are ever reached. It is comprehensive by design, not narrowed to whatever looks like the strongest claim, and it does not analyse any single issue in depth — it spots, flags, and hands off to the skill built for the deeper work.

## Before you start

**The fact pattern itself.** Blocking.

Not blocking, ask once and proceed on a reasonable default without it: **whether issues should be spotted from a specific party's perspective** — what claims a named party could bring — or neutrally across the whole fact pattern. Default to neutral and comprehensive if not specified.

**Governing law or jurisdiction**, if known. Not blocking — work generically if it is not given, but flag every jurisdiction-specific characterisation (a specific cause of action's name or elements) as a verification point rather than asserting it.

## Method

**1. Read the whole fact pattern once before spotting anything.** Issues frequently connect — a limitation problem interacts with when a cause of action is deemed to have accrued, which itself depends on which theory of harm applies — and spotting issue by issue on a first pass misses these connections.

**2. Systematically scan the facts for every legally relevant thread, not just the headline issue.** A contract dispute that also involves personal data, for instance, may raise a data protection issue that a narrower read would miss entirely.

**3. For each issue spotted, name it precisely and tie it to the specific facts that raise it.** A generic label is not useful; the issue needs to be anchored to what actually happened.

**4. Identify potential causes of action for each issue, naming one only where its elements are structurally implicated by the facts as given.** Flag that the exact cause of action's name and elements are jurisdiction-specific and need verification — never assert a specific statute or tort by name with confidence unless it is sourced this session.

**5. Check systematically for threshold problems** — limitation or prescription (flag the specific period as needing verification, never assert it), standing or locus, jurisdiction, and procedural preconditions such as a notice requirement or exhaustion of remedies. These can dispose of an otherwise strong claim entirely, and they are easy to miss when attention is on the substantive merits — treat this as its own deliberate pass, not an afterthought.

**6. Flag issues that are only potentially live depending on a fact not yet established** — "if the contract was signed by an authorised signatory, this; if not, that" — rather than resolving the ambiguity one way. State the conditional plainly.

**7. Group or rank issues by how central they are to what the client actually wants, without discarding the peripheral ones.** A complete issue-spotting exercise stays comprehensive; it does not quietly shrink into a shortlist of the issues that looked most interesting.

**8. Do not analyse any single issue in depth.** This is a spotting exercise. For each issue, name which deeper-analysis skill would take it further — legal-risk-assessor, statutory-interpreter, precedent-mapper, or another — rather than starting that analysis here.

## Output

**1. Header.** Fact pattern summarised, perspective (neutral or party-specific), governing law if given, date.

**2. Issues spotted.** Listed, each tied to the specific facts that raise it.

**3. Potential causes of action.** Per issue, with the reasoning for why its elements are implicated, and jurisdiction-specific naming flagged for verification.

**4. Threshold problems.** Limitation, standing, jurisdiction, procedural preconditions — flagged per issue where relevant.

**5. Conditional or fact-dependent issues.** Flagged, not resolved.

**6. Central versus peripheral.** A ranking or grouping that keeps every issue visible.

**7. Next steps.** Which deeper-analysis skill fits each issue.

**8. Points requiring verification.** Specific limitation periods, cause-of-action elements, and jurisdiction-specific characterisations named but not sourced.

## Do not

Do not analyse any issue in depth. Spot and flag, then hand off.

Do not name a specific statute or cause of action with confidence from memory. Flag it as needing verification unless sourced.

Do not let substantive merits crowd out threshold problems. These are often case-dispositive and are easy to overlook.

Do not resolve a fact-dependent ambiguity. Flag it as conditional.

Do not silently narrow the fact pattern to only the issues that seem strongest. Comprehensiveness is the entire point of this skill.

Referenced files: 1

View saved version →

---
name: legislative-history-analyst
description: Traces how a statutory provision reached its current form — prior versions, amendments, and the reasons for each change — using only sourced material, and keeps an amending act's own stated purpose, a court's later characterisation of that purpose, and this analysis's own inference clearly distinguished as different categories of evidence. Use this whenever a user needs a provision's legislative history traced — including phrasings like "trace how this section reached its current form", "what did this amendment actually change and why", "does the legislative history support reading this provision narrowly", or "was this always the rule or was it added later". Fires for any statutory provision whose history needs tracing, in any jurisdiction — but says plainly when it cannot trace history without sources, rather than reconstructing one from memory.
---

# Legislative History Analyst

## What this does

Traces the history of a statutory provision — its prior versions, each amendment, and the reasons given for each change — using only material actually sourced this session or supplied by the user. Legislative history is exactly the kind of granular factual claim, bill numbers, amendment dates, specific debate quotes, that is easy to produce a convincing-sounding fabrication of; this skill treats every element of the history as something requiring a real source, and says plainly when it cannot trace history at all rather than reconstructing a plausible-sounding account from memory.

## Before you start

**The specific provision and jurisdiction.** Blocking.

**Whether research tools or sources are available this session, or whether the user is supplying historical material.** If neither, say plainly that legislative history cannot be traced without sources, and do not produce an account built from memory. This is the single most important gate in this skill.

Not blocking, ask once and proceed on a reasonable default without it: **how far back the history needs to go** — the original enactment, or a specific set of more recent amendments.

## Method

**1. Confirm the provision and its current text precisely, sourced or supplied.** Do not start from a remembered version of the text.

**2. Identify each prior version or amendment only from sourced material.** State the enactment date, the amending instrument, and precisely what changed in the text between versions — shown as an actual textual comparison, not merely asserted to have changed.

**3. For the reasons behind each change, cite only sourced material** — a statement of objects and reasons, a committee report, a legislative debate record. Do not infer or invent a legislative purpose that is not actually stated in a sourced document.

**4. Keep three categories of evidence for intent clearly distinguished at every point**: the amending act's own stated purpose, a court's later characterisation of that purpose, and this analysis's own inference from the material. Conflating these is one of the most common errors in this kind of analysis, and the output should never let a reader mistake one for another.

**5. Note where the history is genuinely unclear or contested** — conflicting statements about purpose, or gaps in the available record — rather than picking one narrative and presenting it as settled.

**6. Trace amendments in strict chronological order**, and flag explicitly any gap where a version's text is not available or sourced.

**7. Where the user wants the history connected to a current interpretive question, keep the historical account and the interpretive argument visibly separate.** The history is evidence; the interpretation is a conclusion drawn from it, and blending the two overstates what the history alone actually establishes.

## Output

**1. Header.** Provision, jurisdiction, current text (sourced), date, sourcing mode.

**2. Chronological history.** Each version, its amending instrument, date, and what changed, shown as a textual comparison.

**3. Stated reasons for change.** Per amendment, sourced and attributed to its specific source type — statement of objects and reasons, committee report, or debate record.

**4. Gaps and uncertainty.** Any version or reason not sourced, flagged explicitly rather than smoothed over.

**5. Interpretive relevance**, if asked for, kept clearly separate from the historical account itself.

**6. Points requiring verification.** Anything not fully sourced this session.

## Do not

Do not reconstruct a plausible-sounding legislative history from memory when sources are not available. Say so plainly instead.

Do not invent a bill number, amendment date, or debate quote.

Do not conflate an amending act's stated purpose with a court's later characterisation of it, or with this analysis's own inference.

Do not present a contested historical account as settled.

Do not blend the historical account with an interpretive conclusion. Keep them visibly separate.

Referenced files: 1

View saved version →

---
name: precedent-mapper
description: Maps the authorities on a specific legal point and how they relate to each other — which follows which, which distinguishes or overrules another, where a genuine split exists — sourcing every case's existence, citation, and holding only from material retrieved this session or supplied by the user. Use this whenever a user needs the case law on a point actually mapped rather than summarised — including phrasings like "map the authorities on this point", "how do these cases relate to each other", "is there a split in the case law here", or "which of these decisions is still good law". Fires for any point needing its authorities mapped, in any jurisdiction — and says plainly when it cannot map authorities without sources, rather than asserting a case's existence or holding from memory.
---

# Precedent Mapper

## What this does

Maps the authorities bearing on one specific legal point: what each case actually holds on that point, and — the part a simple list does not give you — how the authorities relate to each other, which follows or applies another, which distinguishes it and on what basis, which overrules or has been overruled, and where a genuine unaddressed conflict exists. Case names, citations, and holdings are prime targets for confident-sounding fabrication, so every authority in the map rests on material actually sourced this session or supplied by the user, never on memory.

## Before you start

**The precise point the map is being built around.** A map of "everything about contract law" is not usable; the point needs to be specific enough that an authority either bears on it or does not.

**Whether research tools are available this session, or whether the user is supplying the authorities.** If neither, say plainly that authorities cannot be mapped without sources, and do not produce a map from memory.

Not blocking, ask once and proceed on a reasonable default without it: **how many authorities, or what time period, the map should cover.** An unbounded "map everything" invites an incomplete result presented as though it were complete; bound the scope explicitly.

## Method

**1. Confirm the precise point the map is being built around** before locating a single authority.

**2. Identify authorities only from sourced material — retrieved this session or supplied by the user.** Never assert that a case exists, state its citation, or state its holding from memory.

**3. For each authority, state precisely: its holding on the specific point** — not a general summary of the whole case — **and its precedential weight**: which court decided it, and whether it is binding or merely persuasive relative to the jurisdiction in question.

**4. Map the relationships between authorities explicitly.** Which case follows or applies another; which distinguishes another, and on what stated basis; which overrules or has been overruled; which conflicts with another without either addressing the conflict. This relationship layer is the entire value of a map over a list, and it needs to be stated precisely, never implied or left for the reader to infer.

**5. Identify the current state of the law on the point as it follows from the mapped authorities** — a single clear rule, a split, or an evolving trend — stated as a conclusion that follows from the map, not as an independent assertion sitting alongside it.

**6. Flag where an authority's current validity is uncertain** — potentially overruled by implication, subject to a pending appeal, criticised but not formally overruled — wherever the sourced material actually indicates this.

**7. Note gaps plainly** — a sub-question the mapped authorities do not actually resolve, or a point where research access was limited — rather than letting the map imply a completeness it does not have.

## Output

**1. Header.** The point, jurisdiction, date, sourcing mode.

**2. Authorities.** Each with its citation (sourced), court or level, and its holding on the specific point.

**3. Relationship map.** A table: Authority | Relationship | To which authority | Basis — follows, distinguishes, overrules, or conflicts, stated explicitly.

**4. Current state of the law.** Derived from the map, not asserted independently of it.

**5. Uncertain validity.** Any authority flagged for its current status.

**6. Gaps.** What the map does not resolve, or where research access was limited.

**7. Points requiring verification.** Currency of any authority, and anything not fully sourced this session.

## Do not

Do not assert a case's existence, citation, or holding from memory.

Do not state a general case summary in place of the holding on the specific point being mapped.

Do not imply a relationship between authorities — follows, distinguishes, overrules — without that relationship being sourced.

Do not present the map as exhaustive if research access was limited. Flag the gap.

Do not resolve a genuine split in authority into a false single rule.

Referenced files: 1

View saved version →

---
name: research-synthesiser
description: Weaves research that has already been gathered — from multiple sources, possibly by different people over time, or from other research skills' output — into a single reasoned position on a specific question, surfacing conflicts between sources rather than silently picking a side, flagging gaps rather than filling them, and keeping every step of the reasoning traceable back to the specific supplied material that supports it. Use this whenever a user has scattered research and needs it pulled into one position — including phrasings like "synthesise this research into a single position", "these sources point different directions, help us reason through it", "pull this scattered research together and tell us where we stand", or "what's our overall position given everything we've gathered". Does not conduct fresh research itself — works only from material actually supplied. Fires whenever multiple pieces of existing research need to become one traceable, reasoned answer.
---

# Research Synthesiser

## What this does

Takes research that has already been gathered — from several sources, possibly assembled by different people over time, or the output of skills like precedent-mapper, comparative-analyst, or legislative-history-analyst — and weaves it into a single coherent, reasoned position on a specific question. It does not go and find new material; it works from what is actually supplied, handling conflicts between sources openly and flagging gaps rather than filling them, so that the final position is fully traceable back to the specific material that supports each step of it.

## Before you start

**The scattered research material itself.** Blocking — this is a synthesis skill, not a fresh research skill, and it does not independently gather new material, though it can flag where fresh research would fill an identified gap.

**The specific question the synthesis needs to answer.** Blocking — a synthesis without a target question collapses into a summary, not a reasoned position. State the question precisely before beginning.

Not blocking, ask once and proceed on what is available: **whether new research or verification is available this session to fill a gap identified during synthesis.** If so, use it; if not, flag the gap plainly rather than working around it.

## Method

**1. Read all the supplied research material once before attempting to synthesise anything.** Understand what has actually been gathered before trying to weave it together — synthesising while still reading tends to produce a position that overweights whatever was read first.

**2. Restate the question the synthesis is answering, precisely.**

**3. Map what each piece of supplied research actually establishes.** Do not restate it; extract the specific point it supports, since that is what the synthesis will actually build on.

**4. Identify conflicts between sources explicitly.** Where two pieces of supplied research point in different directions, do not silently pick a side. State the conflict, and reason through it if the material itself supports a resolution — one source more directly on point, more current, or from a higher authority — or flag it as genuinely unresolved if it is not resolvable from what was supplied.

**5. Identify gaps** — what the question needs that the supplied research does not actually cover — and flag these rather than filling them with an assumption or a fact drawn from memory that was not in the supplied material.

**6. Build the reasoned position by showing explicitly how each piece of material supports each step of the reasoning.** A synthesis has to be traceable back to its sources; a freestanding conclusion that merely cites sources afterward is not what this skill produces.

**7. State the confidence of the final position honestly, reflecting the quality and completeness of what was actually supplied.** A synthesis built on thin or conflicting material should say so plainly, not present false confidence to seem more useful.

**8. Where a gap is significant enough that the question cannot actually be answered from the supplied material, say that plainly** rather than stretching thin material to cover a question it does not actually reach.

## Output

**1. Header.** The question, the materials synthesised (listed by source), date.

**2. What each source establishes.** Mapped specifically, not merely restated.

**3. Conflicts identified.** Stated, with reasoning toward a resolution where the material supports one, or flagged as unresolved where it does not.

**4. Gaps identified.** What the question needs that is not covered by the supplied material.

**5. The synthesised position.** Reasoned, with each step traceable to the specific supplied material behind it.

**6. Confidence assessment.** Honest, reflecting the material's actual quality and completeness.

**7. Points requiring further research.** Gaps that need fresh material, naming specifically what is needed.

## Do not

Do not synthesise from material that was not actually supplied. This skill works from what is given; it does not independently research.

Do not silently resolve a conflict between sources without stating it and reasoning through it, or flagging it as unresolved.

Do not fill a gap with an assumption or a fact drawn from memory.

Do not present a synthesis with more confidence than the underlying material actually supports.

Do not produce a position that is not traceable back to the specific supplied material supporting each step of it.

Referenced files: 1

View saved version →

---
name: statutory-interpreter
description: Interprets a specific statutory provision using the recognised canons of interpretation — textual, structural, purposive, and any specific canon whose structural trigger is actually present — checking first whether existing authority already construes the provision, and showing every step of the reasoning rather than asserting a conclusion. Use this whenever a user needs a provision actually interpreted, not just read — including phrasings like "interpret this provision using the accepted canons", "does the ordinary meaning support this reading", "walk through how a court would construe this section", or "is there a stronger interpretive argument for the other reading". Fires for any statutory interpretation question, in any jurisdiction — never assumes a jurisdiction's specific interpretive methodology or canon hierarchy without confirming it.
---

# Statutory Interpreter

## What this does

Interprets a specific statutory provision using the recognised canons of interpretation, showing the reasoning transparently at every step rather than asserting a conclusion and working backward to justify it. It checks first whether existing authority already construes the specific provision — if courts have already interpreted it, that is the primary source and this skill applies it rather than reconstructing an interpretation from first principles as though no authority existed. Every canon applied is named, and applied only where its actual structural trigger is present in this provision, not invoked generically because it is a familiar tool.

## Before you start

**The provision's exact text and the specific interpretive question** — what is actually ambiguous or contested about it, applied to specific facts if relevant. Blocking; interpretation cannot proceed from a paraphrase of the text.

**Governing jurisdiction.** Which canons are recognised, and how they are prioritised relative to each other, varies by legal system — a common-law purposive approach and a civil-law systematic approach are not the same method, and even within common-law systems the accepted hierarchy of canons differs. Do not assume a specific jurisdiction's interpretive methodology without confirming it.

Not blocking, ask once and proceed on what is confirmed: **whether research tools or case law on this specific provision's interpretation are available.** If precedent already construes this exact provision, that is the primary source and should be applied rather than treated as optional context.

## Method

**1. State the provision's exact text and the specific interpretive question precisely** — what is actually ambiguous or contested, not a general description of the topic.

**2. Check first whether existing authority already interprets this specific provision.** If sourced material shows courts have already construed it, apply and report that construction as the primary source rather than reconstructing an interpretation from first principles as though no authority existed. Flag plainly if this check has not actually been done.

**3. Apply the plain or textual meaning canon first.** What does the ordinary meaning of the words yield, including checking any statutory definitions section for terms actually used in this provision.

**4. Apply structural and contextual canons.** How the provision fits within its section, its chapter, and the statute as a whole; whether reading it one way creates surplusage, contradiction, or an absurd result elsewhere in the statute.

**5. Apply purposive interpretation only with a purpose actually sourced** — from a preamble, a statement of objects and reasons, or established case law on the statute's purpose. Do not invent a plausible-sounding purpose that is not actually evidenced.

**6. Apply a specific interpretive canon only where its structural trigger is actually present in this provision** — ejusdem generis where a list is followed by a general term, expressio unius, noscitur a sociis, or another. Name the canon and state precisely why its trigger is present here; do not invoke a canon generically as a rhetorical flourish.

**7. Where legislative history is available, use it as supporting evidence for purpose, kept visibly distinguished from the textual and structural analysis** — or point to legislative-history-analyst if a fuller trace is wanted first.

**8. Show each step of the reasoning in the order it was applied, rather than jumping straight to a conclusion.** The entire value of this skill is the visible reasoning chain; a bare conclusion, however correct, does not deliver it.

**9. State the conclusion with an honest confidence level.** Some interpretive questions have one answer that is clearly correct under the applicable canons; others are genuinely contested. The analysis should reflect which kind of question this actually is.

**10. State the competing interpretation and why the canons favour the preferred reading over it**, rather than presenting only the winning argument as though no other reading was ever plausible.

## Output

**1. Header.** Provision (exact text), the interpretive question, jurisdiction, date, and whether existing authority on this specific provision was checked.

**2. Existing authority**, if found — reported and applied as the primary source.

**3. Textual and plain-meaning analysis.**

**4. Structural and contextual analysis.**

**5. Purposive analysis**, only with a sourced purpose.

**6. Specific canons applied.** Named, with the reasoning for why each one's trigger is present in this provision.

**7. Legislative history**, if used, kept visibly distinct and sourced.

**8. Competing interpretation.** Stated and addressed, not omitted.

**9. Conclusion.** Stated with an honest confidence level.

**10. Points requiring verification.** Anything not sourced — purpose, legislative history, or existing authority not fully checked.

## Do not

Do not skip checking for existing authority interpreting this specific provision before reconstructing an interpretation from first principles.

Do not invent a legislative purpose that is not actually sourced.

Do not invoke an interpretive canon generically without showing why its specific structural trigger is present in this provision.

Do not present only the winning interpretation without addressing the competing one.

Do not assert a specific jurisdiction's interpretive methodology or canon hierarchy without confirming it applies.

Do not jump to a conclusion without showing the reasoning chain that leads to it.

Referenced files: 1