legal-memo
Skill1 starsChanged 35 days agoMITView on GitHub
Write a predictive legal memorandum — Question Presented, Brief Answer, Statement of Facts, an IRAC Discussion (Issue, Rule, Application, Conclusion per issue), and a Conclusion, with Bluebook practitioner citations. Use when the deliverable is an internal legal analysis predicting how a question of law resolves on the facts. Anti-trigger; for a single immutable engineering/technical decision use adr, for a regulated-industry compliance finding rather than a predictive legal question use compliance-audit.
In this repository46 other files it gives its agents.
- academicSkill
- adrSkill
- ai-architecture-docSkill
- arc42-arch-docSkill
- audit-docsSkill
- briefingSkill
- business-planSkill
- c4-model-diagramSkill
- changelogSkill
- clinical-submissionSkill
- competitive-quadrantSkill
- compliance-auditSkill
- computing-paperSkill
- diataxis-explanationSkill
- diataxis-how-toSkill
- diataxis-referenceSkill
- diataxis-tutorialSkill
- doc-set-plannerSkill
- ears-acceptance-criteriaSkill
- engineeringSkill
- exec-summarySkill
- feature-specSkill
- google-design-docSkill
- humanities-chicagoSkill
- humanities-mlaSkill
- kiro-designSkill
- kiro-requirementsSkill
- kiro-tasksSkill
- market-research-reportSkill
- mif-corpusSkill
- mif-frontmatterSkill
- mif-provenanceSkill
- mif-to-pdfSkill
- mif-validateSkill
- nist-spSkill
- playbookSkill
- prdSkill
- python-pepSkill
- regulatory-disclosureSkill
- rust-rfcSkill
- security-pentestSkill
- sre-runbookSkill
- sustainability-reportSkill
- svg-chartsSkill
- systematic-reviewSkill
- trend-analysisSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 53
Add a figure only when authorities or factors are compared on shared attributes: render it as a Markdown table (or a fenced mermaid code …
- Distinctive: line 69
Every claim is traceable to cited authority.
- Distinctive: line 81
Never cite a source you cannot verify is real.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
legal-memo
Produces a predictive legal memorandum: an internal analysis that predicts how a question of law resolves on a given set of facts, and shows its reasoning issue by issue so a supervising attorney or decision-maker can scrutinize the chain from facts to rule to application before relying on it. Its center of gravity is IRAC (Issue, Rule, Application, Conclusion) applied once per discrete issue inside the Discussion — that ordered sub-structure is what makes this a legal memo and not a narrative brief.
This genre follows the standard U.S. law-practice predictive-memo convention, with Bluebook practitioner citation style for authority.
Scope caveat (carry, do not over-sell): this genre reproduces the structure and reasoning form of a predictive legal memorandum. It is not legal advice and asserts no legal sufficiency — the analysis is only as sound as the cited authority it points to.
Audience and altitude
A supervising attorney or internal decision-maker who must act on a prediction of how a legal question resolves. Altitude is predictive: state the most likely outcome and the confidence in it, then earn that prediction issue by issue. Surface adverse authority and counter-arguments honestly — an unaddressed weakness is a defect, not an omission.
Pattern (industry: predictive legal memorandum, IRAC + Bluebook)
- Front matter — a standard memo heading (To / From / Date / Re).
- Question Presented — the legal issue framed as a single yes/no question that joins the governing rule to the determinative facts, not an open-ended topic.
- Brief Answer — the predicted conclusion in 2-3 sentences, with the core reason.
- Statement of Facts — the material facts, stated neutrally; no argument here.
- Discussion — the analytical core, in IRAC per issue: Issue -> Rule -> Application -> Conclusion. One IRAC cycle per discrete issue, each cycle citing its authority. Address counter-arguments within Application. This ordered sub-structure is the genre's distinguishing feature and the section that grounds the Conclusion.
- Conclusion — overall disposition, the recommendation, and any open questions or next steps.
- Back matter — the full table of authorities / reference list; an optional appendix for extended record or verification notes.
Add a figure only when authorities or factors are compared on shared
attributes: render it as a Markdown table (or a fenced mermaid code block if
it is a diagram, never ASCII art or an image), number and caption it, and
reference it in the text. A required figure is never silently omitted — if the
data cannot support one, say so in prose.
Citation style
Bluebook practitioner format — citations to authority as footnotes or inline cites, resolving to a table of authorities in the back matter. Every citation must resolve to a real, verifiable source; no uncited claims. Verify the current Bluebook edition live at authoring time and follow it — never bake an edition number into output as settled fact.
Rules that keep it a legal memo
- Every claim is traceable to cited authority; no orphan facts.
- The Discussion uses IRAC per issue, not a flat narrative — Issue, Rule, Application, and Conclusion must each be identifiable, in order, for every discrete issue.
- State adverse authority and limitations honestly; an undiscussed weakness is a defect, not an omission.
- Hedge uncertain predictions; present the range when authorities disagree, and mark a weakened or contested authority with an explicit qualifier rather than asserting it with the same confidence as settled authority.
- Exhaustive coverage: build the memo from the full set of supplied authority — every surviving source gets its own treatment, never condensed to a cherry-picked subset. A silently dropped source is a defect.
- Never cite a source you cannot verify is real; omit it or substitute a real, well-known authority rather than fabricate a citation.
Anti-triggers — do not use this genre for
- A single already-made technical decision with no predictive legal
question — that is an
adr(Structured MADR): immutable, driver-and-outcome, no IRAC and no legal authority to weigh. - A finding against a regulatory standard rather than a prediction of how a
legal question resolves — that is
compliance-audit: a conformance checklist against a named regulation, not an issue-by-issue legal analysis. - A settled record of what already happened — a legal memo predicts an outcome; it does not narrate history.
MIF frontmatter
type: semantic — a legal memorandum is declarative legal analysis and its
rationale, not a time-bound event or a step sequence. Climb to L2 with
namespace (legal-memo/<area>), modified, title, and tags when the
drafting context supplies them. Gate every output with mif-validate at its
target level; the floor is --level 1.
Why machine-readable — the point of MIF here
| Question an agent asks | Answered by (frontmatter) |
| Is this prediction still current, or has the law moved? | temporal.validFrom / ttl |
| Where did the authority come from; can I trust it? | provenance (W3C-PROV) + citations[] |
| What does this memo relate to or get formalized by? | typed relationships[] (relates-to, realized-by) |
| Which claims are cited authority vs. asserted analysis? | citations[] tied to each IRAC cycle's Rule |
The same document still reads as a human legal memo and projects losslessly to JSON-LD and back — one artifact, two readers.
The L1 -> L3 climb (two exemplars)
templates/good-l1.md— L1 floor:id,type,created+ body. A complete, valid memo, but opaque to a machine consumer.templates/good.md— L3 (highest this genre supports): addsnamespace,modified,temporalvalidity, W3C-PROVprovenance,citations[]tied to the cited authority, and a typedrelationships[]graph (e.g.relates-toa related engineering or compliance record). Validate withmif-validate --level 3.
Author at the highest level the drafting context supports (grade down
rather than fabricate). templates/bad.md shows the antipattern: a memo whose
Discussion has no IRAC sub-structure — the genre's single most load-bearing
required element.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

