Use when a pipeline phase completes and needs an audit trail summary. Called by other skills, not directly by users.
From shieldnpx claudepluginhub infraspecdev/tesseract --plugin shieldThis skill uses the workspace's default tool permissions.
Observes Claude Code sessions via hooks to create atomic project-scoped instincts with confidence scores, evolving them into skills, commands, or agents.
Automatically extracts reusable patterns like error resolutions, workarounds, and debugging techniques from Claude Code sessions via Stop hook, saving them as learned skills for reuse.
Provides patterns for continuous autonomous agent loops with loop selection, quality gates, evals, recovery controls, and failure mitigation. Useful for production AI agent workflows.
Write the summary using the Write tool to exactly this path:
{output_dir}/{feature}/summary/{N}-{slug}/summary.md
Where {output_dir} comes from .shield.json output_dir field (default docs/shield), {feature} is the feature folder name ({feature-name}-YYYYMMDD), {N} is a sequential number, and {slug} is a short kebab-case descriptor (e.g., 1-research, 2-plan-review). Do NOT use any other path, filename, or directory. The Write tool creates directories automatically.
After writing, update {output_dir}/manifest.json and regenerate {output_dir}/index.html.
Called automatically by each phase's orchestrator at the end of execution. Not invoked directly by users.
The orchestrator passes:
phase_name: The pipeline phase that just completed (research, plan, plan-review, sync, implement-step-N, review-step-N, final-review)phase_output: Structured data about what was doneproject_name: From .shield.jsonrun_id: The current run identifier (date-topic){output_dir}/{feature}/summary/{N}-{slug}/summary.md# <Phase Name> Summary
**Run:** <run_id>
**Date:** <timestamp>
## What was done
- Bullet 1
- Bullet 2
## Key decisions
- Decision 1 (if any)
## Findings
- Finding 1 (if any review findings)
## Next
- What the next phase will do