cold-start

Day-one data bootstrapping for a new brain. Sequences the highest-leverage data sources to go from empty brain to useful brain in one session. Uses ClawVisor for safe credential handling — the agent never holds raw API keys. Covers Gmail import, calendar sync, contacts seeding, X/Twitter archive, conversation imports, and file archives. Use when a user has just finished gbrain setup and asks "now what?"

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/cold-start/SKILL.md

Quality

Content

70%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A dense, well-sequenced operational runbook with strong workflow clarity, consent gating, and verification checkpoints. Its weaknesses are repetition of the credential-safety rationale, marketing-style padding around ClawVisor, and a monolithic structure that inlines guide-level detail instead of splitting it into reference files.

Suggestions

State the 'agent never holds raw credentials' rule once in the Contract section and reference it from Phase 0, the decline path, and Anti-Patterns instead of restating the full rationale each time.

Move the full ClawVisor setup walkthrough (dashboard steps, env vars, standing-task scoping) into a references/clawvisor-setup.md and keep only the harness check plus a pointer in SKILL.md.

Replace the prose contacts-CSV and Gmail-mbox parsing instructions with a small script or concrete gbrain commands so every phase is copy-paste executable.

Dimension	Reasoning	Score
Conciseness	The body is mostly actionable (commands, tables, processing rules), but the credential-safety message is repeated three to four times ('The agent never holds raw OAuth tokens or API keys', the Phase 0 'Safety boundary' blockquote, the 'Do NOT hold raw Google tokens yourself' paragraph, and again under Anti-Patterns), and the rhetorical ClawVisor marketing copy ('a live context membrane', 'uncontrolled attack surface') adds padding that could be tightened.	3 / 5
Actionability	Concrete executable bash commands are given for discovery, import, link/timeline extraction, and verification ('gbrain import /path/to/vault --no-embed --workers 4', 'gbrain doctor --json'), and the ClawVisor snippets are explicitly labeled pseudo-code for a specific host class. Minor gaps: contacts CSV parsing and Gmail mbox sampling are described in prose ('Parse the CSV: each row carries...') without a concrete command or script.	4 / 5
Workflow Clarity	Phases 0-8 are clearly sequenced with a priority table, consent gates on every phase, explicit quality gates (sample page after 5 contacts, archive manifest shown before ingestion, approval before '--yes'), verification steps ('gbrain doctor --json', test retrieval queries), a resume protocol keyed on cold-start-state.json, and a per-phase output format. Feedback loops are present for the batch operations, so the destructive/batch cap does not apply.	5 / 5
Progressive Disclosure	There are no bundle files (no references/, scripts/, or assets/), and the single ~550-line SKILL.md inlines content that clearly belongs in separate files — the full ClawVisor onboarding guide (Phase 0), per-source processing rules, and the post-bootstrap checklist. Cross-references to other skills ('skills/google-loops/SKILL.md', 'docs/guides/company-brain-ingestion.md', 'archive-crawler', 'meeting-ingestion') are one level deep and clearly signaled, but the skill itself is a monolith.	3 / 5
	Total	15 / 20 Passed

Description

88%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A well-written third-person description that concretely enumerates its capabilities and gives an explicit, scenario-specific 'Use when' clause. Its only weaknesses are that the description body omits the natural entry phrases (which are relegated to the triggers list) and that the broad source list carries minor overlap risk with dedicated per-source import skills.

Dimension	Reasoning	Score
Specificity	It enumerates six concrete capabilities — 'Gmail import, calendar sync, contacts seeding, X/Twitter archive, conversation imports, and file archives' — plus the sequencing action itself, giving comprehensive concrete coverage rather than generic language. It matches the 'multiple specific concrete actions; comprehensive coverage' anchor and sits clearly above the 'several specific actions; minor gaps' level.	5 / 5
Completeness	The 'what' is explicit ('Sequences the highest-leverage data sources to go from empty brain to useful brain in one session... Covers Gmail import, calendar sync...') and the 'when' is explicit with a concrete trigger phrase ('Use when a user has just finished gbrain setup and asks "now what?"'), matching the anchor that requires both clearly and explicitly with concrete trigger phrases. It is not a 4 because the 'when' clause is already specific rather than merely present.	5 / 5
Trigger Term Quality	Strong natural keywords are present (Gmail, calendar, contacts, X/Twitter archive, conversations, file archives, 'now what?'), matching 'good keyword coverage; a few natural terms missing'. It falls short of the top anchor because the description itself lacks the common entry phrases users would actually say (e.g. 'import my data', 'cold start', 'get started') — those live only in the separate triggers list, not in the description being evaluated.	4 / 5
Distinctiveness Conflict Risk	The niche is clear ('Day-one data bootstrapping for a new brain' tied to gbrain setup completion), but the listed sources (Gmail import, calendar sync, contacts) are domains where dedicated single-source skills plausibly exist, creating minor overlap risk. This fits 'mostly distinct; minor overlap risk with closely related skills' rather than the minimal-conflict top anchor.	4 / 5
	Total	18 / 20 Passed

Validation

87%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (558 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: garrytan/gbrain
Path: skills/cold-start/SKILL.md
Commit: e78f1c3

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.