spectre-capture
Capture durable project decisions, accepted corrections, verified reusable patterns or gotchas, disproved guidance, and changed…
Research a codebase to answer a question — fan out read-only agents (finder/analyst/patterns, web-research for 3rd-party), then synthesize evidence-backed findings into a saved research doc with file:line citations. Use when the user asks \"how does X work\", \"where does Y
$ npx -y skills add joenandez/spectre --skill spectre-research --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/spectre-researchContext preview
The summary Claude sees to decide when to auto-load this skill.
Research a codebase to answer a question — fan out read-only agents (finder/analyst/patterns, web-research for 3rd-party), then synthesize evidence-backed findings into a saved research doc with file:line citations. Use when the user asks \"how does X work\", \"where does Y
name: "spectre-research" description: "Research a codebase to answer a question — fan out read-only agents (finder/analyst/patterns, web-research for 3rd-party), then synthesize evidence-backed findings into a saved research doc with file:line citations. Use when the user asks \"how does X work\", \"where does Y live\", \"investigate/research Z before planning\", or wants a written research artifact. Do NOT use to write code, plan a feature (use spectre-plan), or for a one-off lookup that needs no saved doc." user-invocable: true disable-model-invocation: true
Codebase research: spawn parallel read-only agents, synthesize their findings into one evidence-based document. Live code is the source of truth.
1. **Immediate reply first** — acknowledge the topic and state the feature name/root the workflow will use (or ask for the topic) with NO tool calls in the opening turn. 2. **Read named files fully** in main context (no limit/offset) before decomposing. 3. Decompose the question into areas; track with TodoWrite. Strategy: locate → analyze promising hits → fan out parallel reads. Tell each agent what to find, not how to search. 4. Wait for ALL agents before synthesizing. Compress each return to a 1–2K summary; do not write per-agent scratch files. 5. Synthesize: prefer live-code findings as source of truth; connect findings across components; cite concrete `path:line`; answer the user's actual question with evidence.
Write one research doc to `{FEATURE_ROOT}/research/{topic}_{MMDDYY}.md`; the research doc begins below its title with `Feature: <feature-name>` and `Feature Root: .spectre/features/<feature-name>`. DONE when:
| Handoff | Details | |---|---| | 🧭 **Current phase** | Done | | 📦 **What was just done** | Result | | ▶️ **Proposed next step** | Render resolved action. |
Follow-ups update the same doc (metadata + timestamped section); close with one next-step pointer: Plan with resolved research document path or further research.
SPECTRE is an Agentic Engineering workflow for Claude Code and Codex designed to help you do ONE THING more, faster, and with higher quality.
Capture durable project decisions, accepted corrections, verified reusable patterns or gotchas, disproved guidance, and changed…
Run final adversarial review through a pinned high-effort opposing runtime with one native fallback. Use at the final boundary to…
Generate a grounded draft pull request from the actual diff and open it via gh. Use when wrapping up a branch or writing a PR…
Assess and repair a project's Spectre knowledge, work history, tags, and retrieval quality through primary-led decisions,…
Execute tasks/plans with checks/reviews/proof. Use after planning/resume. Not for planning, unplanned fixes, or pruning.
Internal initializer for a missing Spectre feature root. Invoke only from a standalone workflow with no confirmed managed root.…