From lpm-skills
Produces .docx working views on single legal matters for LPM use, structured by deliverables, issues, risks. Four modes: standard drill-down, decision-first, partner prep, handoff briefing.
npx claudepluginhub legalopsconsulting/lpm-skillsThis skill uses the workspace's default tool permissions.
Working view on a single matter for the LPM's own operational use. Structured around three primary elements: deliverables, issues, and risks. Supporting context orients the reader enough to form questions of others — it does not try to answer every question itself.
Produces portfolio-level morning briefings for legal portfolio managers handling multiple matters, triaging emails, call notes, calendars into meetings, deliverables, issues, and risks as a .docx file. For multi-matter daily sweeps.
Generates legal briefings from email, calendar, contracts, chat, and CRM sources for daily summaries, topic research, or incident response.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Share bugs, ideas, or general feedback.
Working view on a single matter for the LPM's own operational use. Structured around three primary elements: deliverables, issues, and risks. Supporting context orients the reader enough to form questions of others — it does not try to answer every question itself.
The skill is different from status-report-drafter in that the output is for the LPM, not for a partner or client audience. Different register, different structural choices, different expectation of distribution. The drill-down is a working surface; the status report is a communications artefact.
The skill operates in four modes. Standard drill-down (Mode 1) is the default — the LPM has decided to focus on one matter and wants the full working picture. Decision-first (Mode 2) narrows the view to open decisions only. Partner prep (Mode 3) reshapes the template for an upcoming partner call on the matter. Handoff briefing (Mode 4) produces the briefing a receiving LPM needs to take over a matter without losing continuity — a constantly-triggered need that is currently painful to do and frequently skipped despite its operational value.
For portfolio-level views across multiple matters, use daily-briefing. For audience-facing status reports, use status-report-drafter.
Step 1 — Matter count. Scan inputs for matter identifiers, client names, or matter-specific context.
daily-briefing.Step 2 — Mode routing. Identify the user's invocation pattern:
Step 3 — Audience check. If the user specifies the output is for a partner or client audience (e.g. "for JMW to send to the client", "status report for the partner", "client-facing update"), this is the wrong skill. Do not produce any output. Do not produce a status report. Do not produce a client-facing document. Do not read the docx skill. Instead, respond with exactly this pattern and stop:
"This is a [status report / client-facing update] for [audience]. That's status-report-drafter, not a drill-down. Run status-report-drafter on [matter] ([matter number]) with the same inputs."
The failure mode is detecting the audience, deciding you can produce the document anyway, and doing so. That is gate-skipping. The correct behaviour is to refuse and route, even though you could produce the document.
Step 4 — Input tagging. If the user has tagged inputs by source ([FROM LC], [INTERNAL], [CLIENT], [PARTNER]), preserve the tags in attribution.
Matter identifiers. Client name, client number, matter name, matter number. Required in every output header. If not provided in inputs, ask once: "Confirm matter identifiers (client name, client number, matter name, matter number) before I produce the working view." Hard gate — do not produce until confirmed or explicit placeholder agreement given.
Current phase and RAG. Required. If not in inputs, infer from correspondence and flag the inference.
Invocation context. State explicitly why the drill-down is being produced: portfolio briefing handoff, LPMHub click-through, direct ask, upcoming partner call, handoff preparation. Include in output header.
Receiving LPM (Mode 4 only). If Mode 4 invocation, identify the receiving LPM. If not provided: "Who is receiving the matter? (Name required for handoff briefing.)"
Call context (Mode 3 only). If Mode 3 invocation, identify the partner and the call purpose. If not provided: "Which partner and what is the purpose of the call?"
This skill produces a working view on one matter for the LPM's own consumption. It does not:
status-report-drafter)daily-briefing)risk-and-issues-manager)scope-change-controller)budget-and-fee-manager)stakeholder-comms-planner)If the user asks for any of these, route to the appropriate skill and do not produce a drill-down.
Fires when: User requests a single-matter working view without a more specific invocation pattern.
Input: Correspondence, call notes, matter baseline, current-state context for one matter.
Output: Working view using the standard template (below). 2–3 pages.
Fires when: User asks "what do I need to decide on [matter]", "open decisions on [matter]", "[matter] — decisions only".
Input: Same as Mode 1.
Output: Produce a .docx working view using the Mode 2 Template Variation below. Do not produce conversational prose, advisory paragraphs, or a list of decisions as chat output. The output is a structured document with an identifier header, expanded decision entries in the template format, and retained Issues and Risks tables. Deliverables and supporting orientation sections collapsed to one line each. Do not end with a question or offer.
Fires when: User is preparing for a partner call on the matter — "prep me for [partner] call", "before the call with [partner]", "pre-call on [matter]".
Input: Same as Mode 1, plus context on the upcoming call (partner, purpose, timing).
Output: Standard template reshaped for the call:
Fires when: User is handing a matter to another LPM — "handover on [matter]", "I'm handing [matter] to [name]", "leave cover on [matter]", "taking over [matter]", "bringing [name] onto [matter]".
Input: Same as Mode 1, plus identification of the receiving LPM, plus (where possible) the outgoing LPM's knowledge of the matter that is not in the files.
Domain knowledge specific to Mode 4: Handoff briefings are a constantly-triggered operational need. Leave cover, role transitions, and matter reallocations happen continuously. The briefing is universally painful to produce — it takes time the outgoing LPM does not have — and is frequently skipped despite the operational value of doing it well.
The skill's job in Mode 4 is to make the handoff fast enough that it actually happens. The template prioritises the content that matters most to a receiving LPM picking up the matter cold:
Output: Extended template (below). Longer than Mode 1 — 4–6 pages is acceptable because the handoff briefing is the complete operational transfer document.
Produce the template sections in this exact order.
MATTER DRILL-DOWN — [Matter short name]
Client: [Client name] ([Client No.])
Matter: [Matter name] ([Matter No.])
Lead Partner: [Initials or name]
Lead LPM: [Name]
Phase: [Current phase and position]
RAG: [Green / Amber / Red]
Fee model: [Fixed fee / T&M cap / Phased / etc.]
WIP / Budget: [Current WIP] / [Budget]
Prepared: [Date and time]
Invocation: [Why this drill-down is being produced]
SUMMARY
[2–3 sentences. Direct, no hedging. Current position. The one or
two dynamics shaping the week. The next decision point.]
DELIVERABLES — DUE THIS WEEK AND NEXT
| # | Deliverable | Owner | By when | Status |
|---|---|---|---|---|
ISSUES — CURRENT
| Issue | Status | Context for action |
|---|---|---|
RISKS — PLAUSIBLE FUTURE
| Risk | Signal | Mitigation context |
|---|---|---|
OPEN DECISIONS
| # | Decision | Owner | By when | Current state |
|---|---|---|---|---|
SCOPE SIGNALS
Active scope signals on this matter. If none, state "No active
scope signals" and move on.
| Signal type | Driver | Status | Handoff |
|---|---|---|---|
FINANCIAL POSITION
One paragraph. WIP against plan, variance and its nature (structural
vs substantive), forward exposures. Do not reproduce full variance
analysis here — hand off to `budget-and-fee-manager` for detail.
TEAM AND STAKEHOLDER STATE
Internal team: [1–3 lines — who is on, current load, staffing watch items]
Client side: [1–3 lines — who is engaging, current tone, any changes]
External counsel: [1–3 lines — LC status by jurisdiction, escalations]
CROSS-MATTER CONTEXT
How this matter connects to others in the portfolio. Shared partner,
shared LPM, same client, regulatory overlay, counterparty. If no
notable connections, state "No notable cross-matter context" and
move on. Do not invent connections.
HANDOFFS
Single-matter skills to invoke as follow-through.
| Skill | Trigger | Priority |
|---|---|---|
End of drill-down.
Output rule: Produce from available information. Use placeholders for unknowns. Do not withhold pending matter identifiers — ask once, then produce. Do not end with "want me to produce this as a .docx?" — produce the .docx.
Attribution rule: Substantive analysis or recommendations from named individuals are attributed by name. Do not generalise to "the team".
Produce the decision-first working view as a .docx using the template below. Do not produce conversational prose, advisory commentary, or a list of decisions in paragraph form. The output is a structured document with an identifier header, expanded decision entries, and retained Issues and Risks tables. Do not end with a question or offer.
Same header. Collapse Deliverables, Team state, Financial position, Cross-matter context to one line each. Expand:
OPEN DECISIONS — EXPANDED
For each decision:
- Decision statement
- Owner and by-when
- Options under consideration
- Current state (what is known, what is pending)
- What is blocking progress
- LPM recommendation (if applicable)
ISSUES — CURRENT
[Standard table, retained — issues constrain decisions]
RISKS — PLAUSIBLE FUTURE
[Standard table, retained — risks inform decisions]
Same header with Partner, Call purpose, Call timing added.
SUMMARY
[2–3 sentences. Frame specifically around the call.]
DECISIONS REQUIRED FROM THE PARTNER IN THIS CALL
| Decision | Options to consider | LPM recommendation | Why now |
|---|---|---|---|
INFORMATION THE PARTNER MAY NOT HAVE
Items that have come in direct to the LPM, client-side dynamics the
partner has not seen, emerging risks the LPM has spotted. Flagged as
"worth raising".
| Item | Source | Why the partner should know |
|---|---|---|
LIKELY PARTNER QUESTIONS WITH PREPARED ANSWER LINES
3–5 questions the LPM anticipates the partner asking. Each with a
prepared answer line — not a script, a prompt.
| Question | Answer line |
|---|---|
DELIVERABLES — DUE THIS WEEK AND NEXT
[Standard table]
ISSUES — CURRENT
[Standard table]
RISKS — PLAUSIBLE FUTURE
[Standard table]
SUPPORTING CONTEXT
One paragraph each: current financial position, team state, client-side
activity. Enough to orient the partner if they ask.
HANDOFFS
[Standard table]
Same header with Handoff from, Handoff to, Effective date added.
SUMMARY
[3–5 sentences. What this matter is. Where it is today. What is
coming up in the handoff window. Anything the receiving LPM
should know before opening the matter files.]
DELIVERABLES — NEXT 2–4 WEEKS
The operational spine of the handoff. The receiving LPM needs to not
drop any of these.
| # | Deliverable | Owner | By when | Status | Context |
|---|---|---|---|---|---|
ISSUES — CURRENT
Problems in flight that need continued management. Context sufficient
for the receiving LPM to either act or form questions.
| Issue | Status | Mitigation in progress | What is needed next |
|---|---|---|---|
RISKS — PLAUSIBLE FUTURE
Problems that may materialise in the handoff window. Context for
mitigation.
| Risk | Signal | Mitigation context | Watch this specifically |
|---|---|---|---|
OUTSTANDING COMMITMENTS
Commitments the outgoing LPM has made that the receiving LPM must
honour. From emails, calls, meetings — not necessarily documented
in a tracker.
| Commitment | To whom | By when | Context |
|---|---|---|---|
WHAT IS NOT VISIBLE IN THE MATTER FILES
Do not populate this section from inference or from the
correspondence. This section contains only content the outgoing
LPM provides in response to direct prompts. If the outgoing LPM
has not provided this content, replace this entire section with
the prompting questions listed in Domain Knowledge below. Do not
fabricate tacit knowledge. Do not infer relationship dynamics,
communication preferences, or partner behaviour from the tone or
content of emails.
The tacit knowledge. Relationship dynamics, patterns of behaviour,
unwritten context, partner preferences, client quirks. The
outgoing LPM knows these; they are not in the documents.
Examples of what to include:
- How specific client contacts prefer to be communicated with
- Patterns of behaviour on the other side (counterparty, LC, regulator)
- Partner preferences on briefings, updates, and escalation timing
- Client-side politics the receiving LPM should be aware of
- Items the partner knows but that are not in correspondence
WHO TO CALL
By issue type. Name the person, not the role.
| If this comes up | Call | Context |
|---|---|---|
CURRENT FINANCIAL POSITION
Brief. WIP, plan, current variance and its nature, forward
exposures, any fee conversations pending.
STAKEHOLDER TONE SNAPSHOT
Current tone of relationships with key stakeholders. One line each.
- Client GC: [tone and current engagement]
- Client deal lead: [tone and current engagement]
- Lead partner: [posture on matter, any watch items for the LPM relationship]
- Key external counsel: [tone, responsiveness, any issues]
SCHEDULED TOUCHPOINTS — NEXT 4 WEEKS
Calls, meetings, deliverable dates, external milestones.
HANDOFFS
Single-matter skills to invoke as follow-through (e.g. for the
first week of the receiving LPM's tenure).
| Skill | Trigger | Priority |
|---|---|---|
End of handoff briefing.
The output is for the LPM's consumption. No managed-confidence tone. No softening language. No partner-sign-off framing. If the matter is at 91% of cap with Phase 3 not started, say so plainly. If the partner has not delivered on a Friday commitment, say so.
The failure mode is the drill-down drifting into status-report tone. Once it looks like a report, it gets used like a report — circulated, sanitised, partner-reviewed. That is not what this skill produces.
Open decisions are pending judgments — someone needs to decide something. Live items are actions underway — things in motion, deliverables in train.
A drill-down that conflates them loses the critical signal. "Local counsel preparing Swiss supplementary submission" is a live item (deliverable). "Decision whether to split Czech to Phase 3 or re-sequence" is an open decision. The first needs tracking; the second needs someone's attention.
In Modes 1, 3, and 4, both appear but in different sections. In Mode 2, open decisions are the spine.
Same taxonomy as daily-briefing and risk-and-issues-manager:
Silence on its own is a risk. Confirmation promotes it to an issue.
The test for whether a deliverable row earns its place is: does it have a date and an owner? If yes, it is a deliverable. If no, it is either an aspiration or a topic. Deliverables without dates and owners do not appear in the deliverables section — they either get assigned in the output or are moved to Issues/Risks as appropriate.
Section 6 (Financial Position) is one paragraph. WIP against plan, variance and its nature, forward exposures. Do not reproduce detailed variance analysis — the LPM invokes budget-and-fee-manager for that.
Cross-matter context is limited to genuine connections: shared partner, shared LPM, same client, same regulatory overlay, same counterparty. If none present, state "No notable cross-matter context" and move on. Do not manufacture connections. Do not generate cross-matter context from matter-type assumptions or general knowledge. Cross-matter context comes only from input data about other specific matters. If no other matter data is provided, state "No notable cross-matter context" and move on. The fabrication path is inventing plausible connections from the matter type — do not do this.
The hardest and most valuable section of the handoff briefing. Relationship dynamics, unwritten context, patterns of behaviour, tacit knowledge. Prompts for the outgoing LPM:
If the outgoing LPM does not provide this content, the skill prompts for it explicitly before producing Mode 4. This is the one area where a placeholder is not acceptable — the receiving LPM needs the tacit knowledge or the handoff has failed.
Substantive analysis or recommendations from named individuals are attributed by name in the relevant section. Generic attribution to "the team" is the failure mode.
.docx by default. All outputs are .docx unless the user explicitly requests markdown.
Produce, do not ask. Produce on invocation. No conditional offers.
Label is "SUMMARY", not "BLUF".
No preamble. Output opens with the identifier header.
Placeholders for unknowns (except Mode 4 "what is not visible" — see above).
Drill-downs surface matters requiring attorney judgment (legal strategy, client advice framing, privileged communications) but do not offer legal analysis. The skill routes those items to the lead partner for decision via the Partner Attention section — it does not opine.
matter-drill-down is invoked from daily-briefing, from LPMHub, or directly. It hands off to:
scope-change-controllerrisk-and-issues-managerstatus-report-drafterlocal-counsel-managerbudget-and-fee-managerstakeholder-comms-plannertimeline-generatorresource-plannerEach handoff in the output names the trigger and priority.
Mode 1 (Standard drill-down). Connected mode queries the single matter folder in Outlook, retrieves correspondence in a default window (last 7 days unless user specifies), and produces the drill-down without user paste.
Mode 2 (Decision-first). Same retrieval pattern.
Mode 3 (Partner prep). Same retrieval pattern, plus Calendar integration (if available) to identify the last touchpoint with the named partner.
Mode 4 (Handoff briefing). Extended retrieval window — last 30 days minimum — to capture outstanding commitments and pattern signals. The "what is not visible in the files" section still requires the outgoing LPM's input; connected mode cannot generate that content.
Manual mode fallback. All four modes operate fully on pasted input. Connected mode is an efficiency enhancement.
Future connectors. A DMS or practice management system connector would enable matter baseline retrieval (phase, fee model, WIP) from the system of record. A Calendar connector would improve Mode 3 quality by identifying last-partner-touchpoint directly. Both are Phase 3.