rhein1

agoragentic-prime-agent

Govern Prime Agent tool calls with bounded policy, principal-authority checks, redacted evidence, and local receipts. Use before Prime Agent writes, uses networks, deploys, publishes, changes trust, or performs payment-related actions.

rhein1 36 5 Updated 1w ago

Resources

12
GitHub

Install

npx skillscat add rhein1/agoragentic-integrations/agoragentic-prime-agent

Install via the SkillsCat registry.

SKILL.md

Agoragentic for Prime Agent

  1. Start with read-only inspection and no-spend proof.
  2. Classify every proposed tool call as read, write, network, spend, deploy, publish, trust, or unknown.
  3. For spend, deploy, publish, or trust actions, require a short-lived grant bound to the exact principal, agent, session, tool call, capability, and input hash.
  4. Ask interactively for ordinary write/network actions when policy requires review.
  5. Fail closed when review is required but no UI is available.
  6. Require a host-trusted verifier to validate authority integrity; policy allowlists, UI confirmation, and grant fields alone are not authority.
  7. Consume each accepted authority ID and action hash once; retries require a new principal-approved action.
  8. Never let the agent approve its own authority request, expand its own budget, fund its own wallet, or convert a local receipt into settlement proof.
  9. Record hashes and bounded redacted evidence, not raw prompts, credentials, wallet material, or unrestricted tool output.
  10. Reconcile ambiguous paid outcomes before retrying.

A Prime Agent extension is an application policy layer. Prime Agent's worker and kernel processes are not security sandboxes. Payment-bearing and production work still requires a restricted runtime plus enforced network, filesystem, process, and payment chokepoints.

Qualification boundary

  • Bind compatibility claims to Prime Agent v0.7.2, commit 83a0f9f9566219551fcb6ffaf7f519a815749a58, and release SHA-256 bc5471f2a626d727b88a45eb745fff93b10c554a3c4fc5912f25d8c64b987f5e.
  • Verify the raw artifact before extraction; materialize the committed dependency lock in isolation with lifecycle scripts disabled; select the exact Node/platform closure tuple from the schema-closed integrity profile; bind the extracted first-party tree, installed dependency tree, profile, and exact source-extension manifest before spawn.
  • Run with PI_OFFLINE=1, PRIME_AGENT_TELEMETRY=0, --offline, --no-session, --no-builtin-tools, --no-extensions, --no-skills, --no-prompt-templates, --no-themes, and --no-context-files, loading only the exact local extension with -e.
  • Treat the provider-free released-host matrix only as runtime_compatibility evidence. The captured high-severity extract-zip advisory blocks promotion, so keep the candidate level empty and the effective level at source_adapter; do not make a public compatibility claim. The matrix does not prove real policy interception, restricted exact runtime, hosting, production activation, or adoption.
  • Newer-release observation never grants automatic update or promotion authority.
  • Preserve the acyclic manifest → profile → receipt → qualification evidence → Marketplace record chain; verify all runtime-request refs/digests and never feed a downstream hash back into the source manifest.
  • Keep credentials, provider calls, spend, wallet, settlement, deployment, publication, outreach, public compatibility claims, trust, and ranking mutations false, and keep the source-only package centrally held.

Categories