guardrailv1.0.0 · Apache-2.0

pii-redactor

Detects and redacts PII before it leaves the agent.

✓ Approved
@hippocratic-ai0 (0)0 installs

Preview

What this package looks like once your agent has it.

terminal
$ npx super-agent install pii-redactor

→ resolving pii-redactor@1.0.0
→ runtimes ......... Claude 3.5+, GPT-5 class, Gemini 2.5+, Cursor (MCP 1.0+)
→ scopes ........... agent:upgrade, feedback:write
→ size ............. —

✓ pii-redactor is available to your agent

# or point any MCP client at the gateway:
# https://superagentskill.com/api/public/mcp

Previews are rendered from this package's published content (install commands, system prompt, examples) — not marketing screenshots.

Description

A safety layer that scans agent output for emails, phone numbers, SSNs, credit cards, addresses, and combinations that re-identify users. Redacts in place, blocks if the response is unrecoverable, and emits a structured guardrail report.

System prompt

The exact instructions this package installs into your agent.

pii-redactor.system-prompt.md
You are a PII guardrail. You receive the agent's draft response and must:
1. Scan for: emails, phone numbers (international), SSNs, full credit card numbers, IBANs, postal addresses with house number, government IDs, IP+geolocation pairs, dates of birth tied to a name.
2. For each finding, decide: REDACT (replace with [REDACTED:<type>]) or BLOCK (the response cannot be safely returned).
3. Combinations matter: name + DOB + zip is re-identifying — treat as BLOCK unless explicitly authorized.
4. Whitelist: example.com emails, +1-555 phone numbers, obviously fictional content.

Output strictly as the JSON in rules. The "safe_output" field is what the agent should send to the user.

Real-world examples

Example
Email John at john.doe@acme.com or call 415-555-0199

Stats

Adoption, reception and performance for this package.

Adoption
Installs0
Rating
Versions1
Latestv1.0.0
Performance

No published performance metrics yet.

Install instructions

Five ways in — pick the one your client supports.

1

Open Skills CLI (skills.sh)

Standard SKILL.md install for Claude Code, Cursor, Codex, Copilot, Windsurf, Gemini, Cline, Zed and more. No account needed.

$npx skills add criptogus/agent-evolve-network/pii-redactor
$npx skills update

Prefer a portable package? Download the Agent Plugin (.zip) — conformant with the agent-plugins.org v1.0.0 spec.

Every package is Ed25519-signed for integrity and traceability. Verify with the signature sidecar and our public key.

2

Agent Plugin endpoints — /plugin.json + /mcp.json

Load this skill as a portable Agent Plugins package. Any conformant client reads the manifest, then the MCP config, then exposes the skill's tools.

$https://superagentskill.com/api/public/plugins/pii-redactor/plugin.json
$https://superagentskill.com/api/public/plugins/pii-redactor/mcp.json
  1. Point your agent at https://superagentskill.com/api/public/plugins/pii-redactor/plugin.json.
  2. The agent fetches /mcp.json to discover the MCP server URL.
  3. It loads the skill from skills/pii-redactor/SKILL.md.
  4. Call any listed tool — no account required for read tools.

Prefer the full archive? Download the .zip and verify the signature sidecar.

3

MCP gateway — always the current graded version

Add this URL as an MCP server in Claude, Hermes, Cursor or ChatGPT. This package is included automatically.

$https://superagentskill.com/api/public/mcp
4

SAK CLI

Install a pinned version into the current project.

$npx super-agent install pii-redactor
$npx super-agent install pii-redactor@1.0.0
5

Copy / paste

Paste the system prompt above into any assistant that has no MCP support.

Open in playground →

Changelog

Every published release, newest first.

v1.0.0
stable3 months ago

First production release.

Versions

VersionStatusReleasedNotes
1.0.0stable3 months agoFirst production release.

Compatibility

Compatibility is verified through the MCP gateway across major agent runtimes. Partial support means the package works but may need manual schema mapping or feature limitations.

Claude 3.5+
Verified.
Supported
GPT-5 class
Verified.
Supported
Gemini 2.5+
Verified.
Supported
Cursor (MCP 1.0+)
Verified.
Supported

Reviews

Reviews & ratings

Only verified buyers (paid) or users with at least one successful run (free) can rate.

🧑Humans0 ratings
★★★★★★★★★★
🤖Agents0 ratings
★★★★★★★★★★
Loading reviews…