Ask HumanSAFE
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://badge.fury.io/py/ask-human-mcp) [](https://opensource.org/licenses/MIT) [](https://www.python.org/downloads/) [](https://www.coff.ee/masonyarbrough)
stop your ai from hallucinating. gives it an escape route when confused instead of false confidence.
the pain
ai blurts out an endpoint that never existed
the agent makes assumptions that are simply not true and has false confidence
repeat x100 errors and your day is spent debugging false confidence and issues when you could simply ask a question
the fix
an mcp server that lets the agent raise its hand instead of hallucinating. feels like mentoring a sharp intern who actually asks before guessing.
agent → askhuman() ⬇ question lands in askhuman.md ⬇ you swap "PENDING" for the answer ⬇ agent keeps coding
sample file:
### Q8c4f1e2a ts: 2025-01-15 14:30 q: which auth endpoint do we use? ctx: building login form in auth.js answer: PENDING
you drop:
answer: POST /api/v2/auth/login
boom. flow continues and hopefully the issues are solved.
why it's good
- pip install ask-human-mcp → done
- zero config, cross-platform
- watches the file, instant feedback
- multiple agents, no sweat
- locks + limits so nothing catches fire
- full q&a history in markdown (nice paper-trail for debugging)
30-sec setup
pip install ask-human-mcp ask-human-mcp
.cursor/mcp.json:
{
"mcpServers": {
"ask-human": { "command": "ask-human-mcp" }
}
}restart cursor and vibe.
how it works
- ai gets stuck → calls
ask_human(question, context) - question logged → appears
2f0575db89e2OBSERVED · 2026-10-07Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.
claude mcp add ask-human-mcp -- uvx ask-human-mcp
{
"mcpServers": {
"ask-human-mcp": {
"command": "uvx",
"args": [
"ask-human-mcp"
]
}
}
}Exposed tools (3)
3 read · 0 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
ask_human | read | |
get_qa_stats | read | # get some basic stats about the Q&A session |
list_pending_questions | read | # check what questions are still hanging around waiting for answers |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (1)
mcp, watchdog, fastapi, uvicorn, nest-asyncio
Gates applied: no_behavioural_pass.
2f0575db89e2full audit observations/trust-audit/mcp-server/masony817__ask-human.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | 2f0575db89e2 | SAFE | B | 89 | first audit |
Questions
What tools does Ask Human expose?
3 in total: 3 read-only, 0 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is Ask Human safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does Ask Human need?
No credential environment variables were found in its source, so it appears to need none.
How current is this page?
The grade is for one exact copy of the source (2f0575db89e2), read on 2026-10-07. The repository is watched and re-audited when it changes.