agentsclimarketplace

Profile dreamer

Skill yinjialu/context-harness/skills/profile-dreamer

Local-first personal AI context loop for Code Agent conversations

Install
npx -y skills add yinjialu/context-harness --skill profile-dreamer

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Review context-harness conversation archives and propose durable memory/profile updates. Use when the user asks to dream, update memory, refresh their profile, extract personal context from conversations, or maintain the context-harness memory loop.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.1 KB, as published. Nobody here has run it

Profile Dreamer

Use this skill to review archived conversations and propose memory/profile updates. The skill can bootstrap the CLI runtime even when only the skill was installed.

Runtime Bootstrap

Before using local context-harness files or commands, execute the bundled bootstrap script from this skill directory:

runtime_dir="$(bash scripts/bootstrap.sh)"
cd "$runtime_dir"

The script clones or updates the runtime repository at ~/.local/share/context-harness by default, checks out the latest released tag, and runs uv sync.

Overrides:

  • CONTEXT_HARNESS_RUNTIME_DIR: custom runtime checkout path.
  • CONTEXT_HARNESS_REPO_URL: custom fork URL.
  • CONTEXT_HARNESS_REF: custom branch, tag, or commit.
  • CONTEXT_HARNESS_BOOTSTRAP_SKIP_UPDATE=1: skip fetch/pull.

Workflow

  1. Run the Runtime Bootstrap steps above so the local CLI/templates are available.
  2. Resolve the context home in this order:
    • the user's explicit path, if provided
    • CONTEXT_HARNESS_HOME, if set
    • the [paths].context_home value in ~/.context-harness/config.toml, if present
    • otherwise ~/.context-harness
  3. Read:
    • state/dream-state.json
    • global-claude.md
    • memory/user_profile.md
    • new or changed files under conversations/
  4. Extract durable signals:
    • identity changes
    • technical preferences
    • recurring working style
    • active projects
    • writing/content direction
    • explicit instructions for AI agents
  5. Present a reviewable proposal grouped by add/update/strengthen/deprecate.
  6. Wait for user confirmation before writing memory files.
  7. After confirmation, update:
    • global-claude.md
    • memory/user_profile.md
    • logs/dream.md
    • state/dream-state.json

Do not silently rewrite memory/profile files.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.