DOCS/BUILD
GUIDE12 MIN READUPDATED 03 SEP 2026

BUILD / 05

One boundary across the agent stack.

Setup recipes for OpenClaw, Hermes Agent, OpenAI Codex, and Claude Code.
Every runtime launches the same local shreddar-mcp process and can consume the same portable skill. Command and config details differ; the five-tool allowlist and the no-reconstruction boundary do not.
A modular carbon privacy hub accepts context from several connectors, retains coral fragments, and sends lime protected cards into multiple agent modules
SHREDDAR / DOCUMENTATIONOPENCLAW / HERMES / CODEX / CLAUDE / MCP
OPENCLAW

Managed MCP plus workspace skill

TERMINALbash
export SHREDDAR_API_URL=http://127.0.0.1:8787
export SHREDDAR_API_KEY='replace-with-local-gateway-key'

openclaw mcp add shreddar \
  --command /ABSOLUTE/PATH/TO/shreddar/backend/.venv/bin/shreddar-mcp \
  --include 'analyze,protect,shard,get_receipt,get_usage'
openclaw mcp doctor shreddar --probe

mkdir -p ./shreddar-privacy
curl -fsS https://shreddar.pro/integrations/shreddar-privacy/SKILL.md \
  -o ./shreddar-privacy/SKILL.md
openclaw skills install ./shreddar-privacy --as shreddar-privacy
openclaw skills check
HERMES AGENT

Native MCP command plus skill URL

TERMINALbash
export SHREDDAR_API_URL=http://127.0.0.1:8787
export SHREDDAR_API_KEY='replace-with-local-gateway-key'

hermes mcp add shreddar \
  --command /ABSOLUTE/PATH/TO/shreddar/backend/.venv/bin/shreddar-mcp
hermes skills install https://shreddar.pro/integrations/shreddar-privacy/SKILL.md \
  --name shreddar-privacy
hermes mcp test shreddar

# Inside Hermes: /reload-mcp
OPENAI CODEX

Project MCP config plus repo-scoped skill

.CODEX/CONFIG.TOMLtoml
[mcp_servers.shreddar]
command = "/ABSOLUTE/PATH/TO/shreddar/backend/.venv/bin/shreddar-mcp"
env_vars = ["SHREDDAR_API_URL", "SHREDDAR_API_KEY"]
enabled_tools = ["analyze", "protect", "shard", "get_receipt", "get_usage"]
required = true
default_tools_approval_mode = "prompt"
SKILLbash
mkdir -p .agents/skills/shreddar-privacy
curl -fsS https://shreddar.pro/integrations/shreddar-privacy/SKILL.md \
  -o .agents/skills/shreddar-privacy/SKILL.md

# Restart Codex, then inspect /mcp and invoke $shreddar-privacy
CLAUDE CODE

Project-shared MCP plus repo-scoped skill

TERMINALbash
export SHREDDAR_MCP_BIN=/ABSOLUTE/PATH/TO/shreddar/backend/.venv/bin/shreddar-mcp
export SHREDDAR_API_URL=http://127.0.0.1:8787
export SHREDDAR_API_KEY='replace-with-local-gateway-key'

claude mcp add --transport stdio --scope project shreddar -- "$SHREDDAR_MCP_BIN"
mkdir -p .claude/skills/shreddar-privacy
curl -fsS https://shreddar.pro/integrations/shreddar-privacy/SKILL.md \
  -o .claude/skills/shreddar-privacy/SKILL.md

# Inside Claude Code: /mcp then /shreddar-privacy
EVERY RUNTIME

Verify the same invariant

  • Probe the MCP process before the first authorized task.
  • Start a fresh session after installing or changing a skill.
  • Keep SHREDDAR_API_KEY in the runtime environment, never a tracked config value.
  • Test gateway failure and local_only behavior before enabling autonomous tool calls.
SHREDDAR DOCSYOUR DOC. YOUR DATA. YOUR CONTROL.

Public alpha documentation. Capability labels describe implementation state, not certification or guaranteed suitability.