CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

bulk-ingestion

End-to-end discipline for turning any large data source (audio libraries, email takeouts, document corpora, chat exports, API dumps) into brain pages at scale. The lifecycle spine: SCHEMA → ACCESS → TRIAL → EVALUATE → IMPROVE → CODIFY → TEST → SKILLIFY → BULK → MONITOR. State is tracked in a durable JSON manifest (see MANIFEST-PATTERN.md) so any crash, session boundary, or subagent fan-out resumes from ground truth instead of memory.

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./plugin/skills/bulk-ingestion/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

85%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.

An exceptionally well-engineered lifecycle skill: fully actionable commands and templates, a clearly sequenced workflow saturated with validation checkpoints and feedback loops, and disciplined delegation to sibling skills. The two weaknesses are recurring redundancy (repeated manifest references, Contract/dedup overlap) and a broken progressive-disclosure keystone — the heavily-relied-upon MANIFEST-PATTERN.md is not shipped in the bundle.

Suggestions

Ship MANIFEST-PATTERN.md in the skill bundle (or inline its idempotent-worker contract and sharding rules) — the body depends on it for Phases 2 and 9 but the file is missing.

Consolidate the ~8 scattered "see MANIFEST-PATTERN.md" references into one authoritative pointer plus a single 'read before Phase 2' note.

Trim overlap between the Contract section, the Anti-Patterns list, and the phase bodies, which restate the same guarantees three ways.

DimensionReasoningScore

Conciseness

The body is dense and prescriptive with essentially no explanation of concepts Claude already knows; every phase carries rules, paths, or commands. It misses anchor 5 because of real redundancy: the MANIFEST-PATTERN.md pointer repeats roughly eight times, the Contract section restates guarantees proven later in the phases, and "Related skills" re-lists links already given inline.

4 / 5

Actionability

Guidance is executable end to end: a complete copy-ready page-template YAML, concrete commands (`gbrain jobs submit shell --params '...'`, `gbrain agent run "..."`), exact artifact paths (projects/<pipeline-name>/manifest.json, failures.jsonl), separated function responsibilities, and per-error-code remediations (exit 3, exit 11, rate_limited, lock_busy). Placeholders like <your pipeline command> are justified by the generic-pipeline purpose, so there are no meaningful gaps.

5 / 5

Workflow Clarity

Ten phases are explicitly sequenced with a Phases 3-5 quality loop, a 10→100→500→full ramp with quality checks between rungs, artifact-existence verification instead of trusting subagent claims, post-bulk gbrain sync + check-backlinks, a failure log that reopens Phase 5, and a closing Quality Checklist — validation steps, feedback loops, and a checklist are all present, squarely matching anchor 5.

5 / 5

Progressive Disclosure

Section structure is good and references are clearly signaled and one level deep (details delegated to skillify, minion-orchestrator, testing skills rather than inlined). However, the keystone reference MANIFEST-PATTERN.md — cited roughly eight times and flagged as required reading before Phase 2 — is absent from the bundle (no references/ directory and no sibling file), leaving the idempotent-worker contract and sharding details unresolved. That is more than the minor gap of anchor 4, while the structure is far better than anchor 2's monolith.

3 / 5

Total

17

/

20

Passed

Description

66%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 strong, concrete description that clearly communicates a bulk-ingestion lifecycle niche with durable manifest state. Its main weakness is that the explicit 'when to use' trigger guidance (use-when clause, bulk ingest/import phrasing) is split into separate frontmatter fields instead of living in the description text, capping completeness and weakening in-description trigger coverage.

Suggestions

Append a "Use when..." clause with the concrete trigger phrases (e.g., "Use when the user asks to bulk ingest, bulk import, ingest all, or backfill a large source") directly to the description field.

Fold the natural verb phrases from the triggers list ("bulk ingest", "bulk import", "backfill") into the description text so routing works from the description alone.

Briefly boundary the description against type-specific ingestion skills and native gbrain sync to reduce the remaining overlap risk.

DimensionReasoningScore

Specificity

The description names a concrete domain ("turning any large data source (audio libraries, email takeouts, document corpora, chat exports, API dumps) into brain pages at scale") and a 10-stage lifecycle spine, giving several specific elements. It stops short of anchor 5 because the actual capabilities (dedup, entity propagation, trial gating, monitoring) are only implied by phase names rather than listed as concrete actions.

4 / 5

Completeness

The "what" is clear and detailed, but the description text itself contains no "Use when..." clause or equivalent; the 'when' is only weakly implied by "at scale"/"large data source" since the explicit trigger guidance lives in adjacent frontmatter fields (triggers, when_to_use). Per the guideline, a missing use-when clause in the description caps completeness at 3 — not 2 because the 'what' half is explicit.

3 / 5

Trigger Term Quality

Good keyword coverage within the description text — "audio libraries", "email takeouts", "document corpora", "chat exports", "API dumps", "brain pages", "JSON manifest", "subagent fan-out" — comparable to anchor 4's density. The core natural verb phrases ("bulk ingest", "import", "backfill") sit in the separate triggers/when_to_use fields rather than the description itself, so it is not anchor 5's comprehensive coverage.

4 / 5

Distinctiveness Conflict Risk

A clear niche is stated — bulk, manifest-tracked, multi-session ingestion with a distinctive SCHEMA→MONITOR lifecycle — making it mostly distinct. Minor overlap risk remains with type-specific ingestion pipelines and native sync (boundaries are drawn in the body, not the description), so it does not reach anchor 5's minimal-conflict bar.

4 / 5

Total

15

/

20

Passed

Validation

87%

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

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

relative_links

Relative link issues: 5 missing, 12 suspicious

Warning

Total

14

/

16

Passed

Repository

garrytan/gbrain

Path

plugin/skills/bulk-ingestion/SKILL.md

Reviewed

Table of Contents

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.