planning-context

A session-memory guide for repositories that have a `.planning/` directory, where project state is kept in markdown files. It explains how to restore the plan at the start and save progress at the end.

In plain words
What is it for?
Use it to resume work, check the roadmap and current state, update planning documents, and leave a handoff for the next session.
Why use it?
It prevents important project context from being lost when a conversation is cleared or interrupted.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/agenticdevops/aoh/planning-context
Any agent
npx skills add agenticdevops/aoh --skill planning-context
Clone the repo
git clone --depth 1 https://github.com/agenticdevops/aoh

Made for: Claude Code, Codex.

Per session 73 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 521 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00073 $0.00521
Opus 5 $0.00036 $0.00260
Sonnet 5 $0.00015 $0.00104
Haiku 4.5 $0.00007 $0.00052

Measured 2d ago against content hash 01539705193b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

planning-context scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/planning-context/SKILL.md · 45 lines

What it actually says

Planning Context — External Project Memory

Project state lives in .planning/ markdown files, NOT in the context window. Conversation memory is disposable; these files are not. Emulates GSD's .planning/ layout (stay format-compatible) but without GSD process ceremony — superpowers drives the actual work loop.

On session start (or after /clear, compaction, "resume")

Read in order:

  1. .planning/STATE.md — position, next action, session log, handoff notes
  2. .planning/ROADMAP.md — milestones, phases, statuses, definition of done
  3. .planning/PROJECT.md — vision, constraints, decisions table (never re-litigate a decided question; propose a new decision row instead)
  4. As needed: .planning/design/ (dated design docs), .planning/codebase/ (maps), .planning/todos/pending.md (small items)

If chat context and .planning/ disagree, .planning/ wins until the user overrides.

On session close, milestone, or before any risky/large change

Write back BEFORE ending:

  • STATE.md — update Position + Next action; append dated session-log entry (3-6 bullets: what happened, what changed, what's blocked)
  • ROADMAP.md — flip phase statuses; add newly discovered phases to parking lot
  • PROJECT.md — append new decisions: date, decision, why
  • todos/pending.md — add/check off small items
  • Commit .planning/ changes together with the code they describe

Rules

  • Additions over rewrites: append to session log, never erase history
  • Decisions need a "why" — a decision without rationale gets re-litigated
  • Convert relative dates ("yesterday") to absolute (2026-07-14) when writing
  • Keep entries terse and factual; this is memory, not documentation
  • If .planning/ doesn't exist yet, offer to bootstrap it with PROJECT.md, ROADMAP.md, STATE.md seeded from the current conversation
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 45 lines · 73 tokens per session scan A 01539705193b

Subscribe to this mod's changes

planning-context is a skill published in the GitHub repository agenticdevops/aoh (5 stars, last pushed 1mo ago), licensed MIT. It adds 73 tokens to every session and 521 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.