Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add rules/ashlrai/phantom-secrets/cursorrulesgit clone --depth 1 https://github.com/ashlrai/phantom-secretsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00479 | $0.00479 |
| Opus 5 | $0.00239 | $0.00239 |
| Sonnet 5 | $0.00096 | $0.00096 |
| Haiku 4.5 | $0.00048 | $0.00048 |
Grade A, and why
cursorrules scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Phantom Secrets — Cursor Rules
This project uses Phantom Secrets to protect API keys from AI leaks.
Rules
.envcontains phantom tokens (phm_...), NOT real secrets. This is intentional.- Real secrets are in an encrypted vault. Never attempt to read them directly.
- NEVER hardcode API keys in source code.
- NEVER ask the user to paste real API key values into code files.
- NEVER log, display, or write real secret values to any file.
Available Commands
phantom list— See available secrets (never shows values)phantom add <name> <value>— Add a new secretphantom remove <name>— Remove a secretphantom exec -- <command>— Run code with real credentials injected via proxyphantom init— Protect .env secretsphantom rotate— Regenerate all phantom tokensphantom status— Show vault status and secret countphantom doctor— Health check
MCP Server
Phantom has an MCP server with 9 tools. Add it in Cursor Settings > Features > MCP Servers:
- Name:
phantom - Command:
npx phantom-secrets-mcp
MCP Tools
phantom_init— Protect .env secretsphantom_list_secrets— List secret names (never values)phantom_status— Show vault and config statusphantom_add_secret— Add a secret (params: name, value)phantom_remove_secret— Remove a secret (params: name)phantom_rotate— Regenerate phantom tokensphantom_cloud_push— Push encrypted vault to cloudphantom_cloud_pull— Pull vault from cloudphantom_cloud_status— Check cloud sync status
When You See API Keys
If a user shares a real API key (sk-*, ghp_*, xoxb-*, etc.):
- Store it with
phantom add <KEY_NAME> <value>(orphantom_add_secretvia MCP) - Reference the phantom token in code, never the real value
- Never write the real key into any file
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 48 lines · 479 tokens per session scan A 5ca84070cded
cursorrules is a cursor rule published in the GitHub repository ashlrai/phantom-secrets (16 stars, last pushed 13d ago), licensed MIT. It adds 479 tokens to every session, about $0.0024 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other cursor rules, from other repositories
changelog
Update CHANGELOG.md for user-visible changes (Keep a Changelog).
agent-guardrails
把 agent-guardrails 作为 repo-local guardrail 层。.
agent-guardrails-enforce
Before telling the user a task is done, you MUST.
selectools-development
Selectools development conventions, testing requirements, and release process.
selectools-providers
Rules for working with LLM provider adapters in selectools.
pre-action-gates
Before risky tool calls (git push, rm -rf, npm publish, deploy), check prevention rules via the thumbgate MCP server. If a rule matches, STOP and explain why.