adverse-event-narrative
1. Confirm the user objective, required inputs, and non-negotiable constraints before doing detailed work. 2. Validate that the request matches the documented scope and stop early if the task would require unsupported as.
26
17%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./scientific-skills/Academic Writing/adverse-event-narrative/SKILL.mdQuality
Discovery
0%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This description reads like generic process instructions rather than a skill description. It lacks any concrete actions, domain specificity, trigger terms, or 'use when' guidance. Additionally, the text appears truncated ('unsupported as'), further reducing its utility for skill selection.
Suggestions
Identify and state the specific domain or task this skill handles (e.g., 'Validates API request parameters' or 'Scopes data migration tasks') with concrete actions.
Add an explicit 'Use when...' clause with natural trigger terms that a user would actually say when needing this skill.
Remove generic process language ('confirm the user objective', 'validate the request') and replace with specific, distinguishable capabilities that differentiate this skill from others.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description contains no concrete actions or domain-specific capabilities. Phrases like 'confirm the user objective' and 'validate that the request matches the documented scope' are entirely abstract and could apply to any skill. | 1 / 3 |
Completeness | The description fails to answer 'what does this do' in any meaningful way and completely lacks a 'when should Claude use it' clause. It reads like internal process instructions rather than a skill description. | 1 / 3 |
Trigger Term Quality | There are no natural keywords a user would say. Terms like 'non-negotiable constraints', 'documented scope', and 'unsupported as' (which appears truncated) are internal process jargon, not user-facing trigger terms. | 1 / 3 |
Distinctiveness Conflict Risk | The description is entirely generic—confirming objectives and validating scope could apply to virtually any skill. It provides no distinguishing domain, file type, or task category, making it highly prone to conflicts. | 1 / 3 |
Total | 4 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill covers a specialized and important domain (adverse event narrative generation) with genuinely useful regulatory content, quality checklists, and domain-specific pitfalls. However, it suffers from severe verbosity and repetition — the scope description is pasted three times, sections reference each other circularly, and boilerplate sections (error handling, input validation, response template) add significant token cost without proportional value. The core domain content is strong but buried under structural bloat.
Suggestions
Remove all duplicate content: the scope description appears in 'When to Use', 'Key Features', and 'Workflow' — keep it only in the Workflow section and reference it once elsewhere.
Move the detailed CIOMS sections (Core Capabilities subsections 1-4), Quality Checklist, and Common Pitfalls into separate reference files in `references/` and link to them from the SKILL.md overview.
Remove boilerplate sections that teach Claude things it already knows: 'Output Requirements', 'Response Template', 'Input Validation', and 'Error Handling' could be condensed to 3-4 bullet points total.
Add an explicit validation feedback loop in the Workflow: after step 3 (running the script), add a validation step using `validator.py` with a fix-and-retry instruction before proceeding to return results.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose and repetitive. The description string from the frontmatter is copy-pasted verbatim into 'When to Use', 'Key Features', and 'Workflow' sections. There's massive redundancy between sections (e.g., 'Audit-Ready Commands' repeats `--help` twice, 'Implementation Details' says 'See Workflow above' then restates similar content). Many sections explain things Claude already knows (error handling boilerplate, output requirements, response templates). The document could easily be cut by 50%+. | 1 / 3 |
Actionability | The skill provides Python code examples with specific imports and method calls (NarrativeGenerator, timeline analysis, causality assessment), which are concrete. However, these appear to be illustrative rather than verified executable code — the imports reference `scripts.narrative_generator` but no bundle files are provided to confirm these exist. The CLI examples (`python scripts/main.py --help`) are concrete but basic. The quality checklist and common pitfalls sections provide genuinely actionable domain guidance. | 2 / 3 |
Workflow Clarity | The workflow section provides a 5-step sequence with a fallback path, and the 'Example run plan' gives a 4-step execution sequence. However, validation checkpoints are weak — there's a `py_compile` check but no explicit validation feedback loop for the actual narrative output quality. The quality checklist exists but isn't integrated into the workflow steps. For a task involving regulatory submissions (destructive/high-stakes), the lack of explicit validate-fix-retry loops in the workflow caps this at 2. | 2 / 3 |
Progressive Disclosure | The skill references `references/` and `scripts/` directories with clear file listings, and the structure has logical sections. However, no bundle files are provided to verify these references exist. The document itself is monolithic — the extensive CIOMS section details, causality criteria, and format examples could be split into referenced files. The inline content is far too long for a SKILL.md overview, with sections like 'Common Pitfalls' and 'Quality Checklist' that would be better as separate reference documents. | 2 / 3 |
Total | 7 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- aipoch/medical-research-skills
- Commit
73f6514
- 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.