testland/faker-synthetic-data
Substitutes realistic replacement values for PII that a masking or de-identification step removed, nulled, or redacted, so a non-production dataset stays usable. Covers building an injective substitution map that keeps a shared identifier consistent everywhere it appears so joins survive; choosing deterministic (seeded) over random substitution, and the re-identification risk a shared or committed seed reintroduces, since a generator seed is a reproducibility control and not a cryptographic key; preserving field shape where a downstream system validates it, including check-digit values such as payment-card and national-ID numbers plus phone and postal formats; and why a value that merely looks realistic is not yet safe, leaving a residual re-identification measurement over the remaining quasi-identifiers. Use when a masking pipeline has nulled or dropped PII columns and the dataset now needs replacement values that keep cross-table joins intact.
75
94%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Passed
No findings from the security scan
library-surface-and-seeding.mdreferences/
faker-synthetic-data - library surface and seed-exposure detail
Deep reference for faker-synthetic-data. The runnable core - build the substitution map, assert injectivity, preserve check digits, measure residual risk - stays in SKILL.md. This file holds the mechanic tables it draws on.
Minimum library surface
| Mechanic | Behaviour | Source |
|---|---|---|
Faker.seed(n) | Class method; seeds the shared random.Random across all internal generators. Calling .seed() on an instance raises TypeError | faker.readthedocs.io |
fake.seed_instance(n) | Creates and seeds a unique random.Random for one instance | faker.readthedocs.io |
fake.unique.<method>() | Tracks values already returned; raises UniquenessException after repeated failures to find a new one | faker.readthedocs.io |
| Multiple-locale mode | The proxy "randomly select[s] a generator using a distribution defined by the provided weights", so locale varies per call | faker.readthedocs.io |
faker.seed(123) (JS) | Fixes the sequence; setting the seed again resets it. Date helpers also need refDate or faker.setDefaultRefDate() | fakerjs.dev |
To drive substitution from an anonymiser, Microsoft Presidio exposes a custom
operator taking a "lambda to execute on the PII data. The lambda return type
must be a string", passed as OperatorConfig("replace", {"new_value": "BIP"})-style
entries in the operators dict
(presidio.dataprivacystack.org).
Seed exposure matrix
The four seeding strategies referenced from Step 2 of the skill, ranked by the re-identification exposure each one leaves:
| Mechanism | Joins survive | Re-identification exposure |
|---|---|---|
| Seed derived per value, reseeding from the real value before each call | Yes | High. Anyone with the code, the library version, and a candidate list of real values re-runs the derivation and rebuilds the map. The seed reproduces the substitute; it does not conceal the input |
| One global seed plus a stored map | Yes | Equal to the exposure of the map. Control access to the map, not the seed |
| One global seed, map discarded after the run | Yes, within the run | Low from the seed alone: the substitute follows call order, not the input value, so the seed reconstructs nothing without the source data |
| Unseeded, random per occurrence | No | Low, but the dataset is unusable for anything relational |