ai-inventory
EU AI Act per-system inventory — track each AI system's role (provider, deployer, importer,…
First-pass privilege log review — make the obvious privilege calls and flag the hard ones for attorney review without making close calls. Use when the user says "review the privilege log", "priv log", "check privilege on these docs", or has a log to QA before production.
$ npx -y skills add anthropics/claude-for-legal --skill privilege-log-review --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/privilege-log-reviewContext preview
The summary Claude sees to decide when to auto-load this skill.
First-pass privilege log review — make the obvious privilege calls and flag the hard ones for attorney review without making close calls. Use when the user says "review the privilege log", "priv log", "check privilege on these docs", or has a log to QA before production.
name: privilege-log-review description: First-pass privilege log review — make the obvious privilege calls and flag the hard ones for attorney review without making close calls. Use when the user says "review the privilege log", "priv log", "check privilege on these docs", or has a log to QA before production. argument-hint: "[log file, or document set]"
1. Load `~/.claude/plugins/config/claude-for-legal/litigation-legal/CLAUDE.md` → review protocol, priv log format. 2. Follow the workflow and reference below. 3. For each entry: obvious priv / obvious not priv / needs attorney review. Flag reasons. 4. Output: reviewed log with flags. Attorney reviews all flags before production.
---
Before working with a set of litigation documents, ask: "Were any of these documents obtained through disclosure or discovery in legal proceedings?" If yes:
Confirm: "This use is within the proceedings in which the documents were disclosed, or I have permission / consent, or the documents are now public." If not confirmed, flag it: "⚠️ Disclosed documents may have use restrictions. Confirm this use is permitted before proceeding."
**Matter context.** Check `## Matter workspaces` in the practice-level CLAUDE.md. For litigation-legal the default is `Enabled: ✓` — every case gets its own matter workspace. If `Enabled` is `✗` (you turned it off because you work one case at a time), skip the rest of this paragraph and use practice-level context. If enabled and there is no active matter, ask: "Which matter is this for? Run `/litigation-legal:matter-workspace switch <slug>` or say `practice-level`." Load the active matter's `matter.md` for matter-specific context and overrides. Write outputs to the matter folder at `~/.claude/plugins/config/claude-for-legal/litigation-legal/matters/<matter-slug>/`. Never read another matter's files unless `Cross-matter context` is `on`.
---
A privilege log has three kinds of entries: obviously privileged, obviously not, and the ones that need thought. This skill sorts the first two kinds so the attorney's time goes entirely to the third.
**This is first pass. Attorney reviews every flag. No exceptions.**
When this skill cites a rule, local variant, or authority for a privilege call (FRCP 26(b)(5)(A), state rule, local rule, case on waiver scope, case on dominant purpose), two rules apply.
**Pinpoint cites must support the whole proposition.** If the review cites one rule or case to support a multi-part proposition — "the log must describe each document and withhold only materials prepared in anticipation of litigation" — verify the pinpoint covers every element. If it only covers one, split the cite or narrow the proposition. A cite that backs part of a privilege position gets the position rejected when opposing counsel reads the cite and points out it doesn't reach the contested element. This is the "misgrounded citation" failure mode: the cite exists, the passage exists, but it doesn't support the proposition as stated.
**Extract all citations before checking any.** When this review cites authority — or when a separate citation-check is requested on the log, a related brief, or the supporting motion:
1. **First pass: extract.** Read the document and build a list of every citation (rules, cases, statutes, local orders, record cites). Report the count: "Found [N] citations." 2. **Second pass: check.** Check each against the source. Don't sample. Don't stop at the first five. 3. **Report coverage.** "Checked [N] of [M] citations. [K] could not be retrieved — verify manually. [J] confirmed. [I] flagged as potential miscitations. [H] flagged as misgrounded (cite exists but doesn't support the proposition)." 4. **When source text is unavailable, say "could not check," never "confirmed."** A false positive is worse than a "couldn't check" — it lets a bad cite through. 5. **The hardest errors are partial support.** Read the proposition, read the source, compare element by element.
`~/.claude/plugins/config/claude-for-legal/litigation-legal/CLAUDE.md` → privilege log format, review protocol.
**Conflicts gate — unbypassable.** Before reviewing a privilege log, check `~/.claude/plugins/config/claude-for-legal/litigation-legal/matters/_log.yaml` for the matter slug. If the matter is not in `_log.yaml`, refuse and route:
> "I don't see [matter slug] in the matter log. Run `/litigation-legal:matter-intake` first so the conflicts check runs and the matter workspace is set up. I won't review a privilege log on a matter that hasn't been intaken — the conflicts check is the gate, and a privilege log review is work product that needs to live in the matter file."
**Jurisdiction matters.** Privilege scope (A/C and work product), waiver doctrine, and log-form requirements vary materially across federal circuits and state courts. This review applies the rules for the forum specified in config. If the matter involves a different forum, a transferred case, multi-jurisdictional production, or a choice-of-law question on privilege, the calls here may not transfer — re-run against the controlling forum.
**Before reviewing entries, research the forum's
Reference agents, skills, and data connectors for the legal workflows we see most — in-house commercial, privacy, product, corporate, employment, litigation, regulatory, AI governance, IP, and the learning side of the practice (law school clinics and
EU AI Act per-system inventory — track each AI system's role (provider, deployer, importer,…
Run an AI impact assessment — structured intake, risk analysis, regulatory classification per…
Run the cold-start interview — learns your AI governance practice and writes…
Guided customization of your AI governance practice profile — change one thing without…
Manage matter workspaces — new, list, switch, close, or detach (practice-level).…
Keep the AI policy current with practice — weekly sweep of saved AIAs, triage results, and…