.agents/skills/clawdstrike-spec-and-roadmap/SKILL.md
Use when architecture docs exist and need to be hardened into API, storage, migration, protocol, roadmap, or ticket-level specifications for ClawdStrike work.
npx skillsauth add backbay-labs/clawdstrike clawdstrike-spec-and-roadmapInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Use this skill after the architecture is real enough that the next step is not more framing, but sharper specs and executable plans.
Drive toward these outputs:
docs/src/SUMMARY.md only when adding user-facing book docs under docs/src/.*-api-contract.md, *-storage-model.md, *-implementation.md, *-roadmap.md.Read the most relevant existing docs instead of repeating them:
docs/plans/README.mddocs/specs/*.mddocs/roadmaps/*.mddocs/plans/multi-agent/*.md when coordination mattersapps/terminal/docs/dispatch-*.md for the current TUI dispatch initiativeStop when the plan is specific enough that an implementation worker can pick up one slice without another planning roundtrip.
testing
A simple skill demonstrating clawdstrike security
tools
# Hello Skill A simple greeting skill that demonstrates secure agent operation. ## Description This skill allows the agent to greet users and perform basic file operations within the allowed workspace. ## Capabilities - Generate personalized greetings - Read files from the workspace - Write greeting logs to the output directory ## Usage Ask the agent: - "Say hello to Alice" - "Read the welcome message from data/welcome.txt" - "Log a greeting for Bob" ## Examples ### Basic Greeting **Us
testing
# bb-edr: Triage Skill Use this skill to turn clawdstrike audit logs into an incident report and a minimal response plan. ## Inputs - `.hush/audit.jsonl` (JSONL) — clawdstrike audit events (allowed/denied, guard, reason). - `policy.yaml` — the active security policy. ## Task 1. Read and summarize the last ~50 audit events. 2. Focus on **denied** events first: - Group by `guard` (e.g., `forbidden_path`, `egress`, `patch_integrity`) - Identify likely intent (misconfiguration vs. suspici
tools
Threat hunting and security event investigation