BC finance-ocr-pro
Use this skill when the user asks to OCR, transcribe, extract, or convert the contents of a scanned PDF, image, or office document into Markdown, HTML, DOCX, or Excel. This workflow sends page images and OCR prompts to a configured OpenAI-compatible VLM endpoint, requires API_KEY, BASE_URL, and VLM_MODEL, bundles HTML report assets locally so no runtime CDN downloads are needed, and produces improved native Word equation output for recognized LaTeX formulas. It is especially valuable for financial documents and other visually complex materials with dense tables, charts, graphs, and multi-part layouts. Prefer durable background jobs for long-running OCR work.
As a process C 59/100 · Has gaps — weak spots: result and completion, inputs and preconditions, running it twice
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 · 10
✓ No critical or high findings
Medium and low: 10
-
low Obfuscation
obf-base64-blobscripts/md_to_html.py:532Long base64-looking blob (quoted — discussed, not commanded)'/zeO1…H87/976p…zvo+/xXqD/z//XeuR/5vZE/yn/h9h/9nfWk/+n7d/Dz/if/B+'
quoted -
low Obfuscation
obf-base64-blobscripts/md_to_html.py:533Long base64-looking blob (quoted — discussed, not commanded)'53s/eoB///bN6S/wP9ifAt+z/1v8j/3a9k/xH59+uf279ef7d/2ffT/sP6l1jH+L6E/xb61/cP7T/i/9V/dv2k+TP8r9wnp/8uP6'
quoted -
low Obfuscation
obf-base64-blobscripts/md_to_html.py:534Long base64-looking blob (quoted — discussed, not commanded)'r8yPgF/Ff5H/Zf7Z+wn9+/azkM7cegL7K/Ov8D/fv2w/sf7ke4B/Y/3H1J+wn+Y+5v7AP5h/Sv8H/dv23/vP/896z9YvKS+9/9D2'
quoted -
low Obfuscation
obf-base64-blobscripts/md_to_html.py:535Long base64-looking blob (quoted — discussed, not commanded)'Af5H/T/9J/hP3i/xf//+2P+S/5X+O/1n/q/z3uA/Lf8F/wf8j/pP/H/lP//+Bf8m/o3+c/uf+T/6v+M////i+9n//+6T9uv/D7pf'
quoted -
low Obfuscation
obf-base64-blobscripts/md_to_html.py:536Long base64-looking blob (quoted — discussed, not commanded)'7H//Ax7q…SGn'
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/md_to_html.py:539High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)'UxwwM/BYmE…2DK+4BOxJfv//ulPj…OFA/F+WO0f…T5a'
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/md_to_html.py:541High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)'RLoG…0U9+ff5v…k04/3EdNs4+MyNL…Dn7+IZrhr4O'
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/md_to_html.py:543High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)'oHCQ…la5++wkG2…jhF+RlUa…eSX'
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/md_to_html.py:544High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)'wuuH…u5K+quOd/cqOTi+pmGa…dV9/UvfmDNeZFfnGEqMuG/ZGR'
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/md_to_html.py:554High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)'QFkk…dYe+IVjyTTFIPN/Ovw8…7qO/a69kM'
quoted
Files scanned: 22. 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 59/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Progress reporting. Says nothing while it works
- 30Running it twice. 3 mutating operations with no state check
- 60Tools and files. Uses tools (bash, python) that frontmatter does not declare
- 70When it triggers. States when to use, but not when not to
- 100Steps. 36 steps
- 100Failures and branches. 3 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 1275 tokens
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
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
- -310 of 13 scripts are never mentioned in SKILL.md
- +1No license
- +2Single-language instructions
- +3Description length 666: enough signal without eating the budget
- +4Structure: 7 headings
- +3Step-by-step instructions: 36 items
- +4Has examples (6 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 81.