CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

jarvis

Load when acting for the user on their computer, browser, accounts, messages or records, or when starting, handing off or ending a long session.

Quality

64%

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 ./skills/jarvis/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

72%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.

The body is an exemplary lean overview: a routing table plus terse setup rules with concrete probe commands and exit-code checkpoints. Its weaknesses are the bulk destructive cleanup instruction lacking explicit validation, and a large share of referenced paths (all of workflows/, persona.md, README.md) that do not exist in the provided bundle.

Suggestions

Add explicit validation to the Cleanup section, e.g. list each holder's contents and confirm with the user (or a verification command) before removing secrets, databases and records, and define what 'checking the result' means before the backup is deleted.

Ship the referenced files (workflows/session.md, docs-audit.md, efficiency-audit.md, the setup-* guides, persona.md, README.md) in the bundle, or restate their key steps inline so no reading route dangles for a folder-only copy.

Rewrite convoluted sentences like the worktree judgment rule as a short numbered procedure (run, read the line, classify) to lift actionability toward fully executable guidance.

DimensionReasoningScore

Conciseness

The body is dense and un-padded: a routing table, one-line persona rules ("Back up only if a step can lose the only copy; delete the backup after checking the result") and terse setup conditions with no explanations of concepts Claude already knows. Every line carries a distinct instruction, matching 'Lean and efficient; assumes Claude's competence; every token earns its place'.

5 / 5

Actionability

Several instructions are fully concrete and executable, such as the probe "lane-mail events --item overseer --box to-overseer" with exit-code interpretation ("Exit 0 is reachable"), plus "kendex project list" and "kendex refresh". It stops short of 5 because much execution detail is delegated to reference files and some sentences are convoluted, e.g. "judge each folder kendex project list prints, a (worktree of [MAIN]) line by [MAIN]".

4 / 5

Workflow Clarity

The reachable-repositories probe and persona setup have sequences with explicit checkpoints ("Exit 0 is reachable", "Check that the harness loaded those instructions"), but the Cleanup section directs bulk destructive removal ("Remove all unused code, files, keys, secret-store entries, databases, records and settings from every holder") with no explicit validation step beyond 'after checking the result'. Per the destructive/batch cap this cannot exceed 3: 'Steps listed but validation gaps'.

3 / 5

Progressive Disclosure

The routing-table overview pointing one level deep into references/*.md is well organized and clearly signaled, but scored against the actual bundle the navigation breaks: workflows/session.md, workflows/docs-audit.md, workflows/efficiency-audit.md, all four workflows/setup-*.md, persona.md, README.md and ../communication/SKILL.md are referenced but absent, so roughly a third of the reading routes dangle. This is more than 'minor organization gaps', placing it at 'Some structure ... references present but [not resolvable]' rather than the 4 anchor.

3 / 5

Total

15

/

20

Passed

Description

43%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.

The description is a pure trigger clause: it says clearly when to load the skill but never states what the skill does, leaving its capabilities entirely to inference. Trigger terms are natural and reasonably broad, but the missing 'what' and generic action verbs keep specificity and completeness low.

Suggestions

State the 'what' explicitly by leading with concrete capabilities, e.g. "Operates the user's desktop, browser and accounts, drafts and sends messages, and manages long-running sessions" before the 'Load when ...' clause.

Merge the frontmatter summary's concrete verbs ("Guides a delegated desktop assistant through user decisions, checked results and communication") into the description so both what and when are present in one field.

Replace the generic "acting for the user" with 2-3 named actions (e.g. "fill forms", "check results", "reply on the user's behalf") and add common trigger synonyms such as email or files to sharpen specificity and trigger coverage.

DimensionReasoningScore

Specificity

The description names its domains ("computer, browser, accounts, messages or records") but the only actions offered are the generic "acting for the user" and "starting, handing off or ending a long session" — no concrete capability is stated. It matches the anchor 'Names the domain but actions are minimal or generic' and falls below the 3 anchor because it never lists even one or two concrete actions.

2 / 5

Completeness

Only the "when" is explicit ("Load when acting for the user ... or when starting, handing off or ending a long session") while the "what" is entirely absent — nothing states what the skill actually does. This matches the anchor 'only when is present without what' (e.g. "Use when working with documents"); a 3 would require a clear 'what' with weakly implied 'when', which is the inverse of what is written.

2 / 5

Trigger Term Quality

Natural user terms are present — "computer", "browser", "accounts", "messages", "records", "long session", "handing off" — giving good keyword coverage. It sits below the 5 anchor because common variations and synonyms (e.g. email, files, desktop) are missing.

4 / 5

Distinctiveness Conflict Risk

"Acting for the user on their computer, browser, accounts, messages or records" is broad and overlaps browser-automation, credential and messaging skills, though the long-session handoff niche is distinctive. It fits 'Somewhat specific but could still overlap with similar skills'; a 4 would require only minor overlap risk, but "browser" and "accounts" collide directly with dedicated skills.

3 / 5

Total

11

/

20

Passed

Validation

81%

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

Validation — 13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

frontmatter_unknown_keys

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

Warning

relative_links

Relative link issues: 9 missing, 1 suspicious

Warning

Total

13

/

16

Passed

Repository

vanillagreencom/kendex

Path

skills/jarvis/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.