Guide · Agent skill
How to use the Investigation Log Query agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The investigation-query skill is an agent skill that lets an AI assistant answer questions about an open workplace investigation log, such as what witnesses said, where accounts conflict, what evidence is strongest and what is still missing. It is written for employment lawyers and HR investigators working through a large investigation file. After this guide you will know what the skill does, how to install it, how to phrase requests so it triggers, and what it needs to run safely.
Investigation Log Queryanthropics/claude-for-legal · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Download the investigation-query ZIP file from the skill's page on this site.
- 2Install the companion skill. Make sure the internal-investigation reference skill is installed and can be loaded.
- 3Unzip or upload the skill. Unzip into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 4Confirm the investigation log is readable. Make sure the investigation log is in an environment the agent can read.
- 5Ask a question with the matter name. Start with the matter name, then ask a plain-language question such as what witnesses said or where accounts conflict.
- 6Check the citations. Verify each cited log entry ID against the log before relying on the answer.
What an agent skill is and what this one makes the agent do
An agent skill is a folder containing a SKILL.md file that gives an AI assistant specific instructions for a task. When the skill is active and you ask a matching question, the assistant follows those instructions instead of answering from general knowledge.
This skill answers questions against an open investigation log. It can tell you what witnesses said, where accounts conflict, what gaps exist, and what the strongest evidence is on each issue. Each answer cites the specific log entry IDs. If the log has nothing relevant to your question, the assistant says so explicitly and offers to flag it as a gap.
- Summarises what a named person said about an issue
- Compares witness and respondent accounts to find conflicts
- Lists what evidence is missing or still needed
- Cites log entry IDs in every answer
Install the investigation-query skill
Download the ZIP file from the skill's page on this site. Then choose one of the two install routes below.
For Claude Code, unzip the file into ~/.claude/skills/ to make it available to you across projects, or into a project's .claude/skills/ folder to limit it to that project. For the Claude apps, upload the skill under Settings → Capabilities → Skills.
The skill also depends on a companion skill called internal-investigation. Install that too, because the query process and templates live there.
- Claude Code, all projects: unzip into ~/.claude/skills/
- Claude Code, one project only: unzip into the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload
What you need before you start
The skill has two requirements. The companion internal-investigation reference skill must be installed and available to load. Also, an investigation log must be in an environment the agent can read.
The skill runs no scripts and uses no network access, so it works only with the files the agent can already read.
- The internal-investigation reference skill installed and loadable
- An investigation log the agent can read
- No scripts run and no network access used
How to trigger the skill with natural requests
The skill takes a matter name and a question. The skill's argument hint is [matter name] [question]. Name the matter, then ask what you want to know in plain language.
The skill's instructions tell the agent to load internal-investigation and run its Mode 3 (Query) before answering. You do not need to name that step yourself.
/employment-legal:investigation-query [matter name]
What did the respondent say about the December team dinner?Examples of questions to ask
Ask about one person's account, about conflicts between accounts, or about open gaps. Each question below follows the same pattern of matter name first, then the question.
- What did the respondent say about the December team dinner?
- Where do the complainant's and respondent's accounts conflict?
- What do we still need?
/employment-legal:investigation-query [matter name]
Where do the complainant's and respondent's accounts conflict?How to read the answers
Check the log entry IDs in each answer against the log itself. If the assistant reports that it has not seen information on a topic, it will state how many entries it reviewed. You can then ask it to flag that topic as a gap.
Answers only reflect what was recorded in the log. If the log is incomplete or unverified, the answer will be too.
- Verify every cited entry ID
- Treat a 'nothing found' answer as a statement about the log, not about the facts
- Ask the assistant to flag a gap when nothing relevant exists
Safety notes and limits
The skill summarises the log and does not give final legal advice or conclusions. Answers still need professional review.
The log contains sensitive personnel information. Use the skill only in a privileged, confidential setting.
- Not for final legal advice or conclusions
- Not for logs that are incomplete or unverified
- Use only in a privileged, confidential setting
What people use it for
- Finding what a named witness said about a specific incident across a long file
- Checking where a complainant's and respondent's accounts disagree
- Getting a quick list of open gaps before an interview or a report
- Confirming whether the log contains anything about a topic before flagging it as missing
Frequently asked questions
Is the investigation-query skill free?
The source material for this page does not state pricing or licence terms beyond the licence name, which is Apache-2.0. Check the licence file in the source repository for the terms that apply.
Do I need the internal-investigation skill?
Yes. The skill depends on the internal-investigation reference skill for the detailed query process, citation rules and gap-flagging templates. It must be installed and available to load.
Where do I install the investigation-query skill?
Unzip the downloaded ZIP into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder for one project. In the Claude apps, upload it under Settings → Capabilities → Skills.
Does the skill run scripts or use the internet?
No. The skill runs no scripts and uses no network access. It works only with the investigation log the agent can already read.
Can the skill give me legal conclusions?
No. It summarises what is recorded in the log. Its answers still need professional review before you rely on them.
Why does the assistant say it has not seen information on a topic?
If the log contains nothing relevant, the skill is instructed to say so explicitly, state how many entries it reviewed, and offer to flag the topic as a gap.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits employment lawyers and HR investigators who need to query a large investigation log quickly, with citations they can check. It is not suitable for final legal conclusions, incomplete or unverified logs, or any setting outside privileged and confidential handling. If you need a skill that runs scripts or reaches the network, this one does neither, so choose a different tool for those jobs.
Download the skill