fulcra-agent-durable-state

Restore and sync agent tooling across ephemeral machines using the Fulcra File Store.

10|1|Updated May 21, 2026
One-click install
npx skills add https://github.com/ashfulcra/fulcra-tools --skill fulcra-agent-durable-state-ashfulcra
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fulcra-agent-durable-state
Source: https://github.com/ashfulcra/fulcra-tools/tree/main/skills/fulcra-agent-durable-state
Command: npx skills add https://github.com/ashfulcra/fulcra-tools --skill fulcra-agent-durable-state-ashfulcra

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Cloud agent containers get reclaimed and rolled back, silently wiping or reverting the scripts, loops, and config an agent depends on. This Skill treats the Fulcra File Store as the durable home for operational state and local disk as a cache, so agents survive machine loss without from-memory rebuilds or silent downgrades. ## Core Features & Use Cases - Stash lifecycle: Push tooling on change, pull on wake, and list stash contents via the coord-engine stash verbs, with per-file sha256 checksums and executable-bit restoration from a manifest. - Re-entrancy probes: Ordered auth, stash-existence, and local-cache probes tell a waking session exactly where to enter the restore workflow, and every step is safely re-runnable. - Fail-closed secrets guard: stash push refuses secret-shaped filenames (.env, *.key, *token*) and credential-shaped content, keeping shared team paths free of leaked tokens. - Use Case: A fleet coordinator in a rollback-prone container loses its listener loop to a filesystem revert; on the next session it runs the probes, fails the local-cache check, and restores the full bundle with one stash pull instead of rebuilding from memory. ## Quick Start Ask the agent to probe its stash state and restore any missing tooling from the Fulcra File Store before starting work.

Frequently Asked Questions about fulcra-agent-durable-state

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I restore agent tooling after a container rollback?

Run the ordered probes first: check auth with fulcra-api, list the stash with coord-engine stash list, then test each local tool path. If the local cache is incomplete, run coord-engine stash pull with your team, agent name, and destination directory to restore the bundle with checksum verification.

How do I keep agent scripts durable across ephemeral machines?

Push every script or config change immediately with coord-engine stash push, which uploads the file and refreshes the manifest with sha256 checksums. The Fulcra File Store acts as the canonical copy while local disk serves only as a cache.

Can I store credentials or .env files in the stash?

No. Team paths are readable by every agent on the bus, so stash push fails closed on secret-shaped filenames like .env or *.key and on credential-shaped content. Credentials belong in environment configuration, the OS keychain, or an operator-held channel.

What happens if a restored file's checksum does not match?

The stash pull verifies each file's sha256 against the manifest and exits loudly with return code 1 on drift. This prevents silently running a diverged or stale version of a script after a restore.

Do I need coord-engine installed to use the stash?

No. The stash is ordinary files in the Fulcra File Store, so the plain fulcra-api file download and upload commands work as a fallback when the engine is not installed on the host.