Compares git diff against a plan and reports drift percentage by tiered status.
npx skills add https://github.com/linenoize/topia --skill scope-guard-linenoize
Please help me install this Agent Skill. Skill: scope-guard Source: https://github.com/linenoize/topia/tree/main/skills/scope-guard Command: npx skills add https://github.com/linenoize/topia --skill scope-guard-linenoize
This skill prevents scope creep by comparing current changes against the original plan and surfacing unexpected files before they spread into larger work.
.topia/progress.md and extracts intended files, tasks, and exclusions.Use the scope-guard skill to compare the current git diff against the active plan and report any out-of-scope files with drift percentage.
High-intent search queries and answers about installing and using this skill.
Detect scope creep by comparing changed files in your git diff against the original plan, classifying in-scope and out-of-scope changes, and calculating drift percentage for tiered status reporting.
Plan validation prevents drift by loading the active task plan, extracting intended files and exclusions, then inspecting git changes to ensure task execution stays within agreed scope boundaries.
Calculate drift percentage by labeling changed files as in-scope or out-of-scope relative to the original plan, then compute the ratio of out-of-scope changes to assign a tiered escalation status.
Workflow guardrails can flag unplanned features by comparing the current git diff against the active plan, quantifying the drift percentage, and recommending whether to continue, pause, or re-scope.
Scope creep detection requires a pre-existing plan loaded from todo state or progress files to extract intended files and exclusions, without which there is no baseline to classify drift against.
Git diff inspection limitations include reliance on an accurate active plan for baseline comparison, meaning untracked files or unstructured work without explicit task plans cannot be effectively classified for drift.
polseguCompare current implementation and active work against original scope plans.
brainbrewlabsRoutes severity-classified content to appropriate moderation queues or hides it, notifying moderators and updating statuses.
AnveioBuild autonomous coding agents with the Claude Agent SDK for TypeScript.
luanIdentify behavior-preserving simplifications in code diffs with evidence-based verdicts.