ai-disclaimer
Skill63 starsChanged 2 months agoApache-2.0View on GitHub
Write the "AI disclaimer" section for an OsmAnd PR description - a numbered summary of the user's requests in order and what was done for each. Use when asked for an AI disclaimer, PR disclaimer, or a summary of the session for a PR.
In this repository4 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- map/CLAUDE.mdCLAUDE.md
- add-testSkill
- reviewSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 36
When a request exposed a second bug or a wrong assumption, say so in that item, including that the agent asked before touching code if it …
- Distinctive: line 37
Never invent requests, numbers or test results.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
AI disclaimer for a PR
OsmAnd PRs produced with an AI agent end with an ## AI disclaimer section (reference: osmandapp/web#1975). It tells the reviewer how the change came about: which tool, who drove it, and the requests it worked from, in order.
Output
A markdown file AI-disclaimer-<issue or PR number>.md delivered in the chat, for pasting into the PR description. Do not write it into the repo. English, plain prose, no bold, no tables.
Format
## AI disclaimer
Written with <tool> (<model>), driven by [@<github handle>](https://github.com/<github handle>). Summary of the requests it worked from, in order:
1. <request 1, paraphrased>. <What was done or found because of it, one or two sentences with concrete file/function/commit names.>
2. <request 2>. <...>
3. ...
<Closing line: what was verified and by whom, what the agent did not do.>- Tool and model: name what actually ran the session (e.g.
Claude (Cowork, Claude Fable 5.1),Claude Code (Claude Opus 5)), not a guess. - Driver: the user's GitHub handle. The author's handle is
alisa911unless told otherwise.
Rules for the list
- One item per user request, in the order they were made. Paraphrase the request in English; keep the user's intent, including pushback ("Are you sure about the fix?", "Why doesn't any test fail?") - those are part of the process and reviewers want to see them.
- After the request, say what came out of it: the finding, the decision, the change. Name files, functions, ids, commits. This is where agent decisions belong - inside the item that triggered them, not in a separate "Decided by the agent" section. Do not add such a section.
- Merge trivial follow-ups into one item ("Commit name, then a shorter one"). Drop chit-chat entirely.
- When a request exposed a second bug or a wrong assumption, say so in that item, including that the agent asked before touching code if it did.
- Never invent requests, numbers or test results. Numbers in the disclaimer are measurements from real runs; if the user ran the test, say the test was run by the author.
- Closing line states what the agent did not do: no commits made by the agent, tests run by the author, etc.
Length
Usually 5-8 items, each 1-3 sentences. The whole section should fit on one screen.
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.

