Install
$ agentstack add skill-aldianriski-lean-flow-prime ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ● Filesystem access Used
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
prime
Ordered context loader + health check. One-shot session priming. Read-only — never writes.
When to invoke
- First action of a fresh session, before any code-touching work.
- After
/clearto reload context without restarting. - When the session feels stale and you want to confirm what state you are operating against.
Read order
Read each that exists, in order. Mark [OK] / [MISSING] per item. All are optional — absence is reported, never fatal. Canonical placement (DOCS_Guide §2) is listed first; legacy locations second. Adapt the globs to the host project's layout.
| # | Path (first match wins) | Purpose | |---|---|---| | 1 | .claude/CLAUDE.md, CLAUDE.md | Project instructions, anti-patterns, conventions | | 2 | .claude/CONTEXT.md, CONTEXT.md | Vocabulary · gates · modes (if the project uses one) | | 3 | README.md — presence-check by default | Front-door (humans); overlaps 1·2·6 → full read deferred (note below) | | 4 | MEMORY index (harness-resolved; fallback: memory/MEMORY.md, .claude/memory/MEMORY.md) | Sprint state, feedback, references | | 5 | TODO.md, docs/sprint/SPRINT-*.md | Active task list — frontmatter + open - [ ] items only | | 6 | docs/ARCHITECTURE.md, ARCHITECTURE.md | Module map / where-things-live (the durable map) |
README is a fallback (token discipline). It is the human front-door and overlaps CLAUDE.md (1) + CONTEXT.md (2) + ARCHITECTURE.md (6). Stat its presence for the health line, but **read its full content only when CLAUDE.md or CONTEXT.md is MISSING** (then it is the best available overview). Why → docs/research/loop-mechanics-audit.md.
Resolution: read TODO.md (the Backlog pool); follow its § Active Sprint pointer (format: > **SPRINT-NNN — ** → docs/sprint/SPRINT-NNN-.md) — a multi-stream repo lists one pointer per stream — to each active docs/sprint/SPRINT-NNN-.md and read only its frontmatter + Plan (~50 lines). Count open DoD [ ] across all active sprints (report per stream when more than one, e.g. Tasks: 5 open (main: 3 · payments: 2)); if no sprint is active, fall back to the Backlog. Resuming from a /handoff? Also read the handoff doc at the temp path it printed.
Steps
- Read each path in order; track found/missing.
- From the active task list, count open
- [ ]tasks. - Emit the health report (below) — health check ONLY, no inline file summaries.
- Emit one
Next:line:
- open tasks exist →
/orchestratorto continue - no open tasks but a backlog exists →
/lean-doc-generatorto promote / close - nothing tracked yet →
/task-decomposer ""
Output format
=== PRIME HEALTH ===
[OK] CLAUDE.md
[OK] CONTEXT.md
[OK] README.md
[MISSING] MEMORY index
[OK] TODO.md
[OK] ARCHITECTURE.md
Tasks: 3 open
Next: /orchestrator — continue the 3 open tasks
====================
Red flags
❌ Reading files outside the declared slots (beyond a referenced handoff) — adds noise; not this skill's job. ❌ Aborting on a missing file — everything is optional; [MISSING] is the correct response. ❌ Summarizing file contents inline — emit the health check only; files are already in context for downstream skills. ❌ Reading a full sprint file — frontmatter + active section is enough for priming. ❌ Running mid-task — this is for session priming; mid-task use signals context drift.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: aldianriski
- Source: aldianriski/lean-flow
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.