All plugins

scansift

Exact, local triage of SARIF static-analysis results from any SARIF 2.1.0 tool (CodeQL, Semgrep, ruff, Bandit, ESLint, Trivy, Checkov, gosec, Snyk Code, Clang, ...), from one file or a directory of them. sarif_diff compares a branch scan with a baseline scan and returns the new and the gone results, matched by SARIF fingerprints or by content across two checkouts (never by line number), plus results whose level rose, with an optional fail_on gate; sarif_summary gives exact counts by level, the rules and files with the most results, suppressions, security-severity bands and analysis errors; sarif_findings lists results with rule, message, location and snippet, filtered and paged. Standard library only: no network, no credentials, no subprocess (never runs an analyzer), no third-party dependencies, no self-update. Disclosure: reads the local SARIF paths you pass it (read-only), only regular files, and honours Hermes' core read denylist.

rodricksz4h5Toolsv1.0.0Updated 2 days ago

About

Exact, local triage of SARIF static-analysis results from any SARIF 2.1.0 tool (CodeQL, Semgrep, ruff, Bandit, ESLint, Trivy, Checkov, gosec, Snyk Code, Clang, ...), from one file or a directory of them. sarif_diff compares a branch scan with a baseline scan and returns the new and the gone results, matched by SARIF fingerprints or by content across two checkouts (never by line number), plus results whose level rose, with an optional fail_on gate; sarif_summary gives exact counts by level, the rules and files with the most results, suppressions, security-severity bands and analysis errors; sarif_findings lists results with rule, message, location and snippet, filtered and paged. Standard library only: no network, no credentials, no subprocess (never runs an analyzer), no third-party dependencies, no self-update. Disclosure: reads the local SARIF paths you pass it (read-only), only regular files, and honours Hermes' core read denylist.

scansift

Exact, local triage of SARIF static-analysis results, as three Hermes tools.

SARIF 2.1.0 is the common output format of static analyzers and security
scanners: CodeQL, Semgrep, ruff, Bandit, ESLint, Trivy, Checkov, gosec, Snyk
Code, Clang Static Analyzer and many more. A real scan is thousands of results
in a JSON file of many megabytes. ruff over Hermes' agent/, tools/ and
gateway/ with every rule on writes 145 MB and 179,417 results. scansift reads
the file on disk and returns exact answers, and tells you which results a change
introduced.

Why a tool and not a prompt or skill

  • New results, not all results. sarif_diff compares a branch scan with a
    baseline scan. Real output for ruff on two Hermes commits 40 apart: 214 new,
    17 gone, 179,203 unchanged. 204 of the 214 new results are on lines git says
    those commits changed; the rest are whole-function rules reported at a
    definition.

    Results are matched:

    1. by SARIF fingerprints when both scans carry them, in the same file: a full
      fingerprints value in common, or every partialFingerprints key both
      carry agreeing (CodeQL's column fingerprint alone is a small number many
      results share);
    2. otherwise by tool, rule, file, logical location, message (numbers masked)
      and code snippet;
    3. never by line number, which moves with any edit above.

    A result whose level rose (warning to error) is listed under escalated, and
    it counts for the fail_on gate.

    Identical results in one file (thirty "line too long" messages) are paired in
    line order, accounting for how far edits moved the lines around them. That way
    the result reported as new is the one on the edited line, not just any one of
    thirty.

  • Two checkouts, one answer. Analyzers often write absolute paths
    (/tmp/build-4f2/src/app.py). sarif_diff finds the pair of checkout roots
    under which the most files of the two scans line up and compares repository
    paths. Paths outside the checkout (/usr/include/stdio.h) still match as
    they are. Files in different directories are never matched to each other,
    even when they share a name.

  • What the format says, applied exactly:

    • Levels come from the result. Failing that, they come from the
      invocation's ruleConfigurationOverrides, then the rule's default, then
      SARIF's default. A kind other than fail defaults to none.
    • Rules are looked up by index (in the driver or an extension named by index,
      name or guid), then by id, a hierarchical id's parent (CA1000/1), then name.
    • Messages are rebuilt from messageStrings and their {0} arguments, with
      doubled braces read as literal braces. Locations named by artifacts[index]
      are resolved.
    • A suppression without a status, or with status accepted, hides a result;
      underReview and rejected do not.
    • GitHub's security-severity is banded into critical, high, medium and low.
    • An analysis that reports it failed, or that logged errors, is flagged.
      Missing results then are not evidence of clean code.

Tools

All three always return a JSON string. A bad path or a file that is not SARIF
2.x returns { "ok": false, "error": "..." } rather than raising. path
(base/head) is a .sarif / .sarif.json file, or a directory searched for
them. Dotfiles and other JSON are ignored.

sarif_summary

argtyperequireddefaultmeaning
pathstringyes-SARIF file or directory
topintegerno15rules and files to rank (cap 200)

Returns:

  • results, and active (not suppressed, kind fail, open or review);
  • by_level, by_kind and suppressed;
  • security_severity and baseline_state;
  • the rules and files with the most results, worst level first;
  • analysis_errors.

sarif_findings

argtyperequireddefaultmeaning
pathstringyes-SARIF file or directory
rulesstring[]no-only these rule ids
levelsstring[]noallerror, warning, note, none
filesstring[]no-file patterns, * matches any characters
containsstringno-text in the message (case-insensitive)
include_suppressedbooleannofalsealso list suppressed results
offset / limitintegerno0 / 25paging (limit cap 500)

Each result has:

  • rule, rule_name, level, file, line, column and message;
  • snippet, security_severity and help, when the file has them;
  • tool.

Most severe first.

sarif_diff

argtyperequireddefaultmeaning
basestringyes-baseline SARIF file or directory
headstringyes-new SARIF file or directory
fail_onstringno-error, warning or note: passed is false if a new result is at this level or worse
limitintegerno25results per list (cap 500)

Returns:

  • new and gone, each with counts by level and by rule, plus the results
    themselves;
  • unchanged and matched_by;
  • notes when the scans come from different tools or an analysis reported errors.

Suppressed results are left out on both sides.

Limits

  • SARIF 2.x only. SARIF 1.0 (2016) files are refused with a message.
  • Without fingerprints or snippets (ruff writes neither), two identical results
    in one file are told apart only by their line order. That is right for code
    that was edited, but not for code that was moved.
  • A relative URI is taken as the repository path. That is SARIF's
    %SRCROOT% convention. Two scans whose uriBaseId roots point at different
    sub-directories of a repository are not comparable result by result.
  • Results are compared within the same tool. A scan from another tool, or another
    version of a tool with renamed rules, shows as all new and all gone. notes
    says when the tools differ.

Safety

  • Read-only. It reads only the paths you pass: regular files up to 256 MiB, and
    500 files per directory.
  • No network, no credentials, no subprocess (it never runs an analyzer), no
    background work, no self-update. Standard library only.
  • Honours Hermes' core read denylist (credential stores, project .env files,
    NT/device namespace paths). Each path, including each file found in a
    directory, is checked as given and as an absolute path before anything is
    opened. Device, fd and /proc paths are refused, also through symlinks, and
    symlinked directories are not walked.

Install

hermes plugins install rodricksz4h5/scansift --enable

License

MIT

README from the reviewed commit ace9188; it updates when the author re-pins.

More by rodricksz4h5

8 plugins by rodricksz4h5 in the catalog.

rodricksz4h5covsiftExact, local analysis of code coverage reports: LCOV (c8, nyc, Jest, Vitest, cargo-llvm-cov, lcov), Cobertura XML (coverage.py, Jest/Vitest, coverlet, gcovr, PHPUnit), JaCoCo XML, Clover XML, SonarQube generic XML, coverage.py, gcovr and Coveralls JSON, and Go cover profiles, from one file or a directory of them. coverage_diff takes a unified diff (git diff) and returns the added lines no test ran, per file, with diff coverage and an optional fail_under gate; coverage_summary gives exact line, branch and Go statement coverage, the least covered files with their uncovered lines and a per-directory rollup; coverage_file lists one file's uncovered lines and partial branches; coverage_compare shows which files lost coverage between two runs. Standard library only: no network, no credentials, no subprocess (never runs git or tests; the diff is passed in), no third-party dependencies, no self-update; XML entity declarations are refused. Disclosure: reads the local report and diff paths you pass it (read-only), only regular files, and honours Hermes' core read denylist.Communitytoolsrodricksz4h5dataprobeExact, local structure and quality checks over large CSV/TSV and JSONL data files that don't fit in a model's context. data_profile reports per-column inferred type, null rate, distinct count, numeric statistics and top values (for JSONL, the inferred union schema with per-field presence); data_validate reports only the actionable problems — mixed-type columns with the offending line numbers, ragged rows, duplicate rows, non-unique key candidates, constant or all-empty columns and malformed JSONL lines; data_head returns a clean parsed preview that handles quoted CSV and JSON correctly. Streamed, standard library only: no network, no credentials, no third-party dependencies, no self-update. Disclosure: reads the local data path you pass it (read-only) and honours Hermes' core read denylist.Communitytoolsrodricksz4h5layersiftExact, local reading of saved container images: docker save and podman save tars (classic and Docker 25+ layouts) and OCI image archives, including multi-platform indexes and gzip, bzip2 or xz compression. image_summary says where the size goes: each layer with the Dockerfile instruction that made it, the final filesystem size, bytes still shipped after a later layer deletes or overwrites them (which layer added them, which removed them), caches and archives left behind (apt lists, apk/yum/dnf, pip/uv/npm/yarn, Go and Cargo caches, temp directories, .git, compilers) with the fix, and the biggest directories and files; image_layer opens one layer; image_compare compares two images before and after a Dockerfile change. Standard library only: no network, never runs docker or contacts a registry, no credentials, no subprocess, never extracts to disk, no third-party dependencies, no self-update. Disclosure: reads the local image archives you pass it (read-only), only regular files, and honours Hermes' core read denylist.Communitytoolsrodricksz4h5locksiftExact, local review of dependency lockfiles: npm package-lock.json, pnpm-lock.yaml, yarn.lock (v1 and Berry), uv.lock, poetry.lock, Pipfile.lock, Cargo.lock, composer.lock, Gemfile.lock and go.sum. lock_diff compares two versions of a lockfile (a branch against main) and returns the dependency changes instead of thousands of lines of churn, with supply-chain signals per installed copy: a source that moved (registry to git or URL, another registry host, a different tarball or commit at the same version, an alias or patch that changed, a tarball that belongs to another package), checksums changed, removed or weakened, new install scripts, new registries. lock_summary gives counts, packages at several versions, non-registry sources and install scripts; lock_why shows the dependency chains that pull a package in. Standard library only: no network, no credentials, no subprocess (never runs git or a package manager), no third-party dependencies, no self-update. Disclosure: reads the local lockfile paths you pass it (read-only), only regular files, and honours Hermes' core read denylist.Communitytools

Developer Mode

Explore built-in Hermes toolsets and actions.

HermesAgent

Developer?

Create a Hermes extension for Desktop or agent workflows.

Developer Section