Index of every page on this site: llms.txt
All skills
amanktyr avatar

/project-memory

@5711aad
by Aman Ktyramanktyr/tailor6 stars
Run it yourself

Manages the .ai/ progressive project memory layer, context window optimization, document synchronization, and live drift detection between source reality and cached memory.

Use this Skill: https://skilld.dev/gh/amanktyr/tailor/project-memory

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈47 tokens for metadata: the name and description. ≈566 when used: this file.

Description uses 2.4% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈189 characters, or 2.4%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Project Memory Skill

Activate this skill when reading, creating, updating, or repairing .ai/ memory files, resolving documentation drift, or managing agent context windows.

Operational Principle: Project memory is an indexed cache of project facts. Source code is the ultimate ground truth.


1. Project Memory Hierarchy & Source of Truth

When conflicting information is detected between files:

  1. Source of Truth 1 (Highest): Verifiable source code, database schemas (schema.prisma, models.py, migrations/), and package manifests (package.json, Cargo.toml, go.mod).
  2. Source of Truth 2: Project Memory Files (.ai/STACK.md, .ai/API.md, .ai/DATABASE.md).
  3. Source of Truth 3 (Lowest): Outdated inline comments, legacy markdown notes, or unverified chat instructions.

If memory conflicts with source code: Trust the source code, flag the drift, and immediately update the memory document.


2. Progressive Disclosure Rules (Token Optimization)

To minimize context token consumption and maintain maximum LLM reasoning capacity:

  • Entry Point: Always read .ai/CONSTITUTION.md and .ai/INDEX.md first.
  • Domain Documents (Read strictly on demand):
    • Architecture decisions: .ai/ARCHITECTURE.md, .ai/DECISIONS.md
    • UI components and design tokens: .ai/UI.md, .ai/CONVENTIONS.md
    • Endpoints, contracts, and payloads: .ai/API.md
    • Schemas, migrations, and ORMs: .ai/DATABASE.md
    • Installed packages and governance: .ai/DEPENDENCIES.md
    • Security policies and secret management: .ai/SECURITY.md
    • Verification runners and commands: .ai/TESTING.md

3. Drift Detection & Live Repair Workflow

When finishing any non-trivial coding task or executing /update-project-memory:

  1. Check if dependencies were added, removed, or upgraded. If so, synchronize .ai/DEPENDENCIES.md.
  2. Check if new database models or migrations were introduced. If so, synchronize .ai/DATABASE.md.
  3. Check if new API endpoints or route contracts were created. If so, synchronize .ai/API.md.
  4. Check if new architectural decisions were adopted. If so, record an ADR in .ai/DECISIONS.md.
  5. Verify that .ai/INDEX.md accurately references all active domain memory files.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts3mo3 checks · Risk SAFE
  • Gen Agent Trust Hub3mo

    This skill manages project documentation and metadata within the `.ai/` folder, ensuring consistency between source code and project memory. No security risks were identified.

  • Socket3mo

    No alerts

  • Snyk3mo

    Risk: LOW · No issues

Signed by skilld at 5711aad. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 3 months ago

README badge

README badge for amanktyr/tailor/project-memory