Skip to content

Repo Build & Repair Tasks

Coding tasks from written specs and real issues, each with a repository that runs and hidden tests that grade the change.

  • 96,000repo tasks delivered
  • 9,600spec-to-repo tasks delivered
01

Sample

2 records

Grading contract for one environment

Real field shape from a shipped task. Identity, symbol names, and mutation rules are withheld here on purpose - publishing them would contaminate a graded environment we sell.

meta.json
{
  "task_id": "05-<withheld>",
  "kind": "mutated-domain",
  "import_name": "<withheld>",
  "public_symbols": ["<5 renamed symbols>"],
  "grader": {
    "pytest_paths": ["."],
    "min_passed": 193,
    "source": "Regenerated from the authored reference. Asserts the MUTATED semantics."
  },
  "reference_install": "reference",
  "known_bad_install": "known-bad",
  "upstream_sdist": "<real upstream tarball, plagiarist gate only>",
  "anti_memorization": "Public surface renamed; every rule inverted from the well-known
    analogue. A memorized reproduction fails the structural check first and the
    behavioral long tail second - the public package, installed as is, fails the task.",
  "provenance": "Authored from scratch. Upstream is retained only to prove it fails."
}

Held-out assertion distribution across the reference tranche

Real per-task counts from the 12-task reference tranche. Task identities are withheld.

assertions.csv
task,origin,held_out_assertions,plagiarist_gate
01,invented,178,none
02,mutated,121,upstream pinned
03,mutated,74,upstream pinned
04,mutated,63,upstream pinned
05,mutated,193,upstream pinned
06,mutated,149,upstream pinned
07,mutated,51,upstream pinned
08,mutated,61,upstream pinned
09,mutated,30,upstream pinned
10,invented,161,none
11,invented,116,none
12,invented,111,none
,,1308,8 of 12 gated
02

Specifications

Repo Build & Repair Tasks specifications
FormatsDocker, JSONL, Patch, YAML
DeliverySecure download, Docker image, Git bundle
CadenceContinuous build; monthly tranches
Spec-to-repo tasks delivered9,600
Long-horizon repo tasks delivered96,000
Held-out assertions30 to 193 per task in the reference tranche
Environmentpip-installed in Docker, history-stripped
GradingFAIL_TO_PASS + full PASS_TO_PASS regression
AccessOrg-scoped credentials with per-customer delivery paths. Held-out graders ship under a separate restricted agreement from the solver-visible specifications so evaluation tasks can be split from training tasks at delivery.
Built againstNL2Repo-Bench (ICML 2026), SWE-bench Verified, SWE-bench Pro

What ships

  • Spec-to-Repo Tasks

    A natural-language library specification plus a held-out assertion suite. The agent writes the library; the suite decides. Every task is validated against a plagiarist gate that proves the answer is not reachable by copying an existing package.

  • Long-Horizon Repository Tasks

    Multi-file, cross-subsystem changes mined from real permissively-licensed repositories - real issue, real fix, real held-out tests - served in a history-stripped environment so the answer cannot be read out of the Git log.

  • Teacher Trajectories

    Full tool-call traces from frontier coding agents solving these environments, with reasoning preserved, graded on the same held-out suites and passed through a Git-history cheat scan.

Fields

11 fields

FieldDescription
task_idstringStable task identifier within the tranche.Example 07-<withheld>
kindenuminvented-domain | mutated-domain | mined-issue.Example mutated-domain
import_namestringModule the structural check imports before any behavioral test runs.
public_symbolsstring[]Symbols the candidate must expose; a memorized upstream package fails here first.
grader.min_passedint · assertionsExact held-out assertion count, counted from the regenerated suite.Example 193
grader.pytest_pathsstring[]Held-out suite roots. Never shipped inside the solver-visible tree.
reference_installpathAuthored correct implementation. Must install and pass at the ship bar.
known_bad_installpathPlausible-but-wrong package. Must install cleanly, then fail the suite.
upstream_sdiststring · nullableReal upstream tarball for the plagiarist gate; null for invented domains.
anti_memorizationstringWritten record of every renamed symbol and mutated rule, kept private to the grading tree.
provenancestringAuthorship and licensing of the reference implementation.

Load it

The first commands after an approved delivery.

quickstart.sh
# Grade a candidate against one held-out environment
bash harness/grade_in_docker.sh \
  --task <task-id> \
  --install ./candidate      # path-only; bare requirement strings are refused

# The runner executes with the network namespace off:
#   docker run --network none ...      pip install --no-index ...
# Score = passed_assertions / min_passed        PASS only at the full ship bar

bash harness/verify_all.sh   # reference PASS / known-bad FAIL / upstream FAIL
03

Quality checks

4 checks

Plagiarist gate

Whether reproducing the real public library passes the task.

The genuine upstream distribution is installed into the task environment and run through the task own grader. The gate requires it to fail.

Real measured: upstream fails on every mutated-domain task in the reference tranche. The four invented-domain tasks have no public analogue to gate against.

Known-bad discrimination

Whether the grader scores behavior rather than mere presence of the right symbols.

A plausible-but-wrong package is required to pass the structural symbol check and then fail the behavioral suite.

Real measured: passes on every task in the reference tranche.

Held-out assertion count

How much graded surface each task actually carries.

Assertions are counted directly from the regenerated suite and recorded as the exact pass threshold in the grading contract.

Real measured: 30 to 193 held-out assertions per task across the 12-task reference tranche.

Copy guard

Whether a submission is byte-near the shipped reference or the real upstream source.

A similarity check runs on every submission with a configurable hard-fail threshold, as defense in depth behind the network isolation.

Methodology-stage. Reported per submission; no aggregate false-positive rate is published.

Ground truth

What is correctThe authored reference implementation and its regenerated held-out suite. For mined repository tasks, the real upstream fix commit and the tests that commit made pass are the ground truth.
How it is establishedA structural symbol check runs first, then the held-out suite. Score is the fraction of held-out assertions passed, so an honest 80 percent implementation scores 0.80 rather than zero. A task only counts as passed at the full ship bar. Repository tasks additionally require the complete pre-existing suite to stay green.
AgreementGrading is deterministic and execution-based, so there is no inter-rater statistic. Discrimination is evidenced instead by the known-bad and upstream gates, both of which are required to fail.

How it is built

  1. 01

    Author or mutate the domain

    Each task is either an invented domain with no public analogue, or a familiar domain whose public surface is renamed and whose semantics are deliberately inverted. A model that reproduces the library it remembers is wrong on the graded long tail rather than merely unlucky.

  2. 02

    Regenerate the held-out suite from the reference

    Tests are generated from the authored reference implementation and assert the mutated behavior. No upstream test suite is ever vendored in, so the suite cannot agree with a memorized package by construction.

  3. 03

    Structurally isolate the held-out material

    The solver receives exactly one file per task: the specification. Reference, regenerated tests, known-bad package, grading contract, and upstream tarballs live in a sibling private tree, and the verifier asserts that nothing but the specification exists in the solver-visible directory.

  4. 04

    Grade offline and path-only

    Grading runs in a pinned image with the network namespace off. Installs use a no-index flag and bare requirement strings are refused outright, so the grader can only ever score source the candidate wrote.

  5. 05

    Prove the grader discriminates

    Every task is verified four ways before it ships: the reference passes at the ship bar, the known-bad package installs and then fails, the real upstream library run through the task grader fails, and the solver-visible tree contains nothing but specifications.

  6. 06

    Mine and validate long-horizon tasks

    Repository tasks are mined by locating paired source and test commits in permissively-licensed projects, then validating each as a genuine fail-then-pass transition inside a history-stripped environment. Candidates that do not reproduce that transition are discarded.

05RequestReplies usually within one working day

Request a sample

We send real records in the format you would receive, with the field list and where each record came from. Access for evaluation is limited and covered by a mutual NDA.

ProductRepo Build & Repair Tasks

team@gerra.com