Estimate implementation effort for proposed engineering work by comparing anticipated change scope against repository size and structure. Use when triaging tickets, sizing backlog items, deciding whether to decompose work, deciding whether to stop and plan before coding, or producing pre-implementation LoE. Produces deterministic story-point estimates, lines changed, files touched, decomposition recommendations, planning recommendations, and blast-radius risk guidance.
repoRoot for the target repository.baseRef and headRef when an actual diff exists.Skill-local resources:
scripts/estimate_execution.py and references/estimation-rubric.md relative to this
SKILL.md.python3 <skill-dir>/scripts/estimate_execution.py \
--repo-root /path/to/repo \
--base-ref origin/main \
--head-ref HEAD
python3 <skill-dir>/scripts/estimate_execution.py \
--repo-root /path/to/repo \
--proposed-files /path/to/proposed-files.txt
estimation.storyPoints: recommended LoE in story points.change.filesTouched: expected/observed files touched.change.linesChanged: expected/observed line churn.comparison: relative impact versus repository baseline.risk.blastRadius.score and risk.blastRadius.level: deterministic blast-radius assessment.risk.blastRadius.recommendedControls: stricter review and test controls for high-risk work.risk.blastRadius.investigationAreas: adjacent areas to inspect when the touched paths imply wider impact.estimation.decompositionRecommended: whether to split the work item.planning.recommended: binary guidance on whether to stop and plan before coding.planning.matchedRules and planning.rationale: explicit rules that explain why the recommendation is yes or no.decompositionRecommended is true.risk.blastRadius.requiresHeightenedControls is true, explicitly add broader regression coverage, adjacent-boundary review, and the listed investigation items before execution or merge.planning.recommended is true, stop after estimation and present a concrete execution plan before implementation.planning.recommended is false, proceed directly unless the user explicitly asks for a plan.Return a concise summary plus the JSON artifact fields:
scripts/estimate_execution.py relative to this skill directoryscripts/blast_radius.py relative to this skill directoryscripts/planning_recommendation.py relative to this skill directoryreferences/estimation-rubric.md relative to this skill directory