Menu
Iteratively enhance the forge container and environment by running diagnostics, building, and committing changes autonomously.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Discover, claim, implement, checkpoint, and complete units of work from the shared plan ledger based on host capabilities and lease rules, complementary to multihost-orchestration.
Build, install, and DESTRUCTIVELY smoke-test the current Tillandsias checkout on the host you are running on — Linux (Podman), macOS (Virtualization.framework VM), or Windows (WSL2). Detects the OS, builds + installs the locally built tray, irreversibly destroys the host's runtime substrate (Podman store / macOS VM dir / WSL2 distro), re-provisions from a pristine state, and — on Linux only — launches the forge with `/meta-orchestration`. Every issue observed is filed as a `plan/issues/` work packet.
Coordinate Tillandsias Linux, Windows, and macOS implementation agents by auditing shared plan/methodology ledgers, analyzing sibling branch git history, reconciling stale work queues, mediating concurrent conflicts or thrashing, tracking convergence velocity metrics, enforcing finite-time convergence guarantees, and pushing coordination updates. Designed to run hourly to ensure continuous cooperation and unblocking.
Host-aware Tillandsias recurring runtime loop: sync remote state, drain claimable plan work, run eligible e2e smoke gates, coordinate integrations on mutable Linux, release when warranted, update plan, commit, and push before exit.
Clean-room end-to-end smoke test of a PUBLISHED release. Curl-installs the latest release binary from GitHub, does a full `podman system reset --force`, runs `tillandsias --debug --init` from a pristine state, and (if init is clean) launches `tillandsias . --opencode --prompt "Use the /meta-orchestration skill"`. Every issue observed — by this skill's agent during install/reset/init, or by the agents running inside the forge under meta-orchestration — is filed as a plan/issues work packet for later pickup by /advance-work-from-plan.
Refresh linux-next, open/update the PR to main, merge once green, bump VERSION, tag, and trigger the workflow_dispatch release. Runs on a daily cadence so the latest linux-next work reaches a downloadable Linux release operators can smoke-test on a Fedora Silverblue host.
| name | forge-continuous-enhancement |
| description | Iteratively enhance the forge container and environment by running diagnostics, building, and committing changes autonomously. |
Purpose: This skill runs inside the forge to improve the forge itself iteratively. By capitalizing on the safe YOLO-mode environment (highly permissive settings for opencode and codex/claude), agents can confidently inspect diagnostics, shape proposals, claim approved plan packets, test build outputs, measure telemetry, and push durable plan or implementation checkpoints.
Workflow:
tillandsias codebase (e.g., via tillandsias --opencode --prompt "Use the /forge-continuous-enhancement skill").Containerfile builds (e.g. build-install.log or telemetry logs) to identify warnings, errors, and unoptimized layers../plan/issues/ or plan/forge-improvements/proposals/./advance-work-from-plan discipline before editing implementation files.main.Pre-requisites: