BC research-assistant
Enrich Bear research notes tagged 「待整理」 with theme-matched GIFs. Use when the user wants to auto-illustrate pending research notes, batch-process Bear notes for visual enrichment, or automate the "find a fitting GIF" step of note cleanup. Requires Bear + grizzly CLI.
As a process C 52/100 · Has gaps — weak spots: result and completion, failures and branches, consistency
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".
If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 4
✓ No critical or high findings
Medium and low: 4
-
medium Exfiltration
net-credential-usescripts/enrich_note.sh:35Credential used in a network call (verify the destination is the intended service)GIF_URL=$(curl -sf "https://api.giphy.com/v1/gifs/search?q=…&api_key=…&limit=1&rating=g" \
-
low Exfiltration
exfil-secret-in-urlscripts/enrich_note.sh:30Credential passed in a URL query string (normal for some APIs — verify the host is the intended service) (placeholder value)GIF_URL=$(curl -sf "https://tenor.googleapis.com/v2/search?q=…&key=…&limit=1&media_filter=gif" \
placeholder -
low Exfiltration
net-credential-usescripts/enrich_note.sh:30Credential used in a network call (verify the destination is the intended service) (destination is a well-known publishing service)GIF_URL=$(curl -sf "https://tenor.googleapis.com/v2/search?q=…&key=…&limit=1&media_filter=gif" \
known service -
low Exfiltration
exfil-secret-in-urlscripts/enrich_note.sh:35Credential passed in a URL query string (normal for some APIs — verify the host is the intended service) (placeholder value)GIF_URL=$(curl -sf "https://api.giphy.com/v1/gifs/search?q=…&api_key=…&limit=1&rating=g" \
placeholder
Files scanned: 4. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
✓ No remarks against the Agent Skills spec
Process rating: all ten parameters 52/100
- 0Result and completion. Does not say what the result is
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 30Running it twice. 3 mutating operations with no state check
- 40Consistency. Frontmatter name (research-assistant) differs from the folder (bear-gif-enricher)
- 60Tools and files. Uses tools (bash, web) that frontmatter does not declare
- 70When it triggers. States when to use, but not when not to
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 14 steps
- 100Execution cost. Instruction body is 268 tokens
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Output format is not stated: the model decides each time
- +4No input/output examples
- +1No license
- +2Single-language instructions
- +3Description length 267: enough signal without eating the budget
- +4Structure: 4 headings
- +3Step-by-step instructions: 14 items
- +3All 2 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 83.