Skill · v1.0
Pending Plan Implementation
Pending Plan Implementation is a skill for executing approved plans with checkpoint gates, breadcrumb logging and outcome summaries. Each step has explicit approval boundaries. The plan note becomes the implementation context, the execution log and the outcome record in one document.
Install
Claude Code (CLI / WSL / Git Bash)
/plugin marketplace add https://www.infinitegameos.io/marketplace.json
/plugin install pending-plan-implementation@igos-libraryClaude Code (VS Code)
Install in VS CodeOpens the Claude Code plugins dialog with the marketplace and skill prefilled. Requires the Claude Code VS Code extension installed and signed in. Or paste the snippet below into .claude/settings.json for VS Code, JetBrains or any setup that prefers manual config.
{
"extraKnownMarketplaces": {
"igos-library": {
"source": {
"source": "url",
"url": "https://www.infinitegameos.io/marketplace.json"
}
}
},
"enabledPlugins": {
"pending-plan-implementation@igos-library": true
}
}Direct markdown URL (Claude Code, Cursor, Codex CLI)
https://www.infinitegameos.io/markdown/skills/pending-plan-implementationCursor (.mdc rules file)
curl -O https://www.infinitegameos.io/install/cursor/pending-plan-implementation.mdcAider, Cline, any agent with --read
curl -O https://www.infinitegameos.io/markdown/skills/pending-plan-implementation
aider --read pending-plan-implementation.mdDefinition
Pending Plan Implementation executes an approved plan (or approved subset) using the plan note itself as the implementation context, breadcrumb record and outcome summary. The skill confirms intent and risk before starting, documents the execution scope and what remains out of scope, executes approved work step by step, logs every action in the plan's execution log, updates evidence references, proposes status changes (approval-gated) and reads back the implementation summary before applying any approval-gated changes. An optional second-AI audit gate offers a challenger pass on substantive output documents before close, with a one-line opt-in framing.
Execution flow with approval gates
The skill opens by reading back plan intent, risk, affected areas and current breadcrumbs. The operator confirms exactly what is approved to implement now and what remains out of scope. The skill then executes step by step, logging actions in the plan's execution log section as it goes. Status changes never auto-apply; they surface as proposals for explicit consent.
An optional audit gate at Step 7 assesses whether the primary output document would benefit from a second-AI challenger pass. A one-line opt-in offers it without requiring it. If yes, the configured second-AI tool reviews the document inline; findings work through in the session, and durable insights graduate to a Refinements section. The audit is a session event, not a separate ritual.
Use Cases
Multi-session implementation with state preservation
A plan runs across three sessions. Each session updates the plan's execution log with what shipped, evidence references and outcome notes. The next session resumes from the plan note itself; no separate continuity doc required. The plan IS the breadcrumb record.
Approval-bounded subset execution
An approved plan has eight steps. The operator approves only steps 1-3 for this session. The skill executes 1-3, logs them, surfaces what remains out of scope and proposes the implementation_state advance. Steps 4-8 wait for the next approval window.
Substantive output with second-AI audit
A session produces a new codex section. The optional audit gate offers a challenger pass before close. The operator accepts; the second-AI tool reviews the section. One finding warrants a permanent rule. It graduates to the Refinements section as a single date-stamped entry.
FAQ
How is this different from just executing a plan manually?
Manual execution often skips the breadcrumb update, leaves the plan note stale and loses track of what shipped vs what remains. This skill makes the plan note the source of truth: every action logs there, every status change proposes there, every outcome records there. State preservation is structural, not incidental.
What is the audit gate for?
It is a one-line opt-in offered before session close on substantive output documents (codex, protocol, governance artifact, finished argument). The operator can accept, decline or skip. When accepted, it routes the document to a second-AI tool for adversarial review. Optional, never default.
Can it auto-execute approved plans?
No. The skill confirms approved scope explicitly at the start of each session. Status changes propose for explicit consent. Approval is granular and per-session, not blanket and persistent.
Pending Plan Implementation pairs naturally with Plan Challenger. Challenge first, then implement. Both are tools for an operator playing a longer game. The full architectural frame that motivates per-plan governance is the Sovereign Life Playbook.
See the Sovereign Life Playbook