feat(security): config-driven lifts of 6 skills (PR2/5)#386
Merged
potiuk merged 1 commit intoMay 30, 2026
Conversation
Second of 5 PRs converting the security skill family from Airflow/ASF-coupled to a generic framework. PR1 (apache#381) landed the schema + adapter contracts with ASF defaults. This PR lifts 6 skills to read those knobs (and the existing sibling adopter-config files) instead of inlining ASF/Airflow values. Byte-equivalent for the airflow-s adopter: every value the skill currently inlines either becomes a reference to a config knob whose ASF default matches today's inlined value, OR keeps the inlined airflow-s value as a single named example in generic prose. Per-skill lifts: - security-tracker-stats-dashboard (+15/-6 SKILL.md, +14/-2 default-config.yaml) — cross-reference `scope_detection.labels` and `tracker.labels` in default-config.yaml comments; no literal default values changed. - security-issue-deduplicate (+26/-15) — golden-rule scope cross- merge, scope-check, milestone shapes, and the CVE-record URL now reference `scope_detection.labels`, `<project-config>/milestones.md`, and `cve_authority.record_url_template`. - security-issue-import-from-md (+18/-5) — `<security-list>` placeholder uses, body-field heading mapping -> `tracker.body_fields`, label list -> `tracker.labels`, scope rule -> `scope_detection.labels`. - security-issue-fix (+78/-48) — toolchain commands (`uv`, `breeze`, `prek`) -> `<project-config>/fix-workflow.md`; package registries -> `release_process.artifact_registries`; `apache/airflow` -> `<upstream>`; `main` -> `<default-branch>`; local-clone probe paths kept as airflow-s examples in generic prose. - security-issue-triage (+48/-27) — scope-label triads -> `scope_detection.labels`; canned-response examples reframed as airflow-s named examples in generic guidance; `@`-handle routing -> roster references. - security-issue-import-from-pr (+85/-45) — biggest lift in this batch. Project-board node IDs de-inlined (already in project.md "GitHub project board" table). Scope cascade (`airflow|providers|chart` triad) -> `scope_detection.labels` with the airflow-s mapping kept as one named example. `Apache Airflow:` title-prefix -> `<vendor>: <product>:` derived from `project.md`. Aggregate: +280/-129 lines across 7 files. Validator clean (1 pre-existing soft warning on an unrelated skill). 218 tests green. No new placeholders introduced beyond those declared in PR1's schema. Out of scope: no ASF-default adapter (`tools/vulnogram/`, `tools/ponymail/`, `tools/gmail/asf-relay.md`) is touched — those are PR3/PR4. No skill outside the 6 above is touched. The deeper Vulnogram-state-machine refactor in `security-issue-sync`, `security-cve-allocate`, and `security-issue-invalidate` is PR4. The forwarder-relay sub-skill extract is PR3. Generated-by: Claude Code (Opus 4.7)
5 tasks
potiuk
added a commit
that referenced
this pull request
May 30, 2026
Third of 5 PRs converting the security skill family from Airflow/ASF-coupled to a generic framework. PR1 (#381) landed the adapter contracts. PR2 (#386) lifted the config-driven skills. This PR extracts the ASF-Security forwarder + PonyMail archive specifics into: - a new optional sub-skill `security-issue-import-via-forwarder` that the generic intake/invalidate/sync skills invoke when `forwarders.enabled` is non-empty - explicit cross-references from the adapter-contract READMEs to their default-ASF implementations (`tools/gmail/asf-relay.md` for the asf-security forwarder; `tools/ponymail/` for the ponymail mail-archive) Byte-equivalent for the airflow-s adopter: every behaviour the ASF-relay row + Step 5d + Step 2b previously produced is now reachable via the sub-skill, which the airflow-s adopter installs by default (`forwarders.enabled: [asf-security]`). == New sub-skill == `.claude/skills/security-issue-import-via-forwarder/SKILL.md` (+620 lines). Adapter-agnostic body — no `asf-security` / `huntr-relay` / `hackerone-relay` string in control flow. Reads enabled adapters from `forwarders.enabled` in project.md, dispatches via `detect()` / `extract_credit()` / `reporter_addressing_block()` from `tools/forwarder-relay/README.md`. Four steps: Step 0 — Pre-flight check Step 1 — Detect adapter match Step 2 — Extract reporter credit Step 3 — Route reporter-facing drafts Step 4 — Hand back to parent skill Frontmatter: `capability: capability:intake`. Validator clean. == Skill lifts == - security-issue-import (-60/+54 net -6) — dropped the ASF-security-relay row from the Step 3 classification table; replaced with a pre-classification paragraph that points at the sub-skill when `forwarders.enabled` is non-empty. Generalized the Step 7 receipt-of-confirmation routing and every other inline `ASF-security relay` / `Report` pairing (golden-rule prose, Step 4 field-extraction header, Step 5 proposal grouping, Step 6 default-disposition, rollup provenance template, Hard Rules). - security-issue-invalidate (-49/+93 +44) — Step 5d ASF-relay inline logic replaced with adapter-aware routing through the sub-skill + `tools/forwarder-relay/README.md`. Four touch- points lifted; ASF retained as a named example in worked- example sections. - security-issue-sync (-20/+20 ±0) — scoped Step 2b lift only (the big Vulnogram-state-machine rewrite at Steps 5b/5c is PR4). Draft routing now reads adapter metadata from the sub-skill's hand-back; no inline `Dear PMC` preamble match. == Adapter-contract README cross-references == - tools/forwarder-relay/README.md (+24) — explicit "Implementation: tools/gmail/asf-relay.md" pointer for the asf-security adapter + sub-skill consumer link. - tools/mail-archive/README.md (+19) — explicit "Implementation: tools/ponymail/" pointer + the skills that consume PonyMail today (intake / sync / invalidate). == Doc table == docs/labels-and-capabilities.md gets a new row for `security-issue-import-via-forwarder` → `capability:intake` (satisfies the capability-sync check). Aggregate: 7 files changed, +795/-116. Validator clean (5 advisory soft warnings, none hard, none on PR1/PR2-touched files). 218 tests green. Out of scope (deferred to PR4/PR5): - `tools/vulnogram/`, `tools/gmail/asf-relay.md` bodies (this PR only updates cross-references TO them, not their content) - `tools/ponymail/` body (same) - `security-issue-sync` Steps 5b/5c CVE-state-machine rewrite (PR4 — the ~600-line section) - `security-cve-allocate` Vulnogram-specific body (PR4) - `docs/security/process.md`, `forwarder-routing-policy.md`, `roles.md`, `threat-model.md` (PR5) Generated-by: Claude Code (Opus 4.7)
5 tasks
potiuk
added a commit
that referenced
this pull request
May 30, 2026
Fifth and final PR of the security genericization series. Lifts the remaining 4 docs in docs/security/ to read config knobs from projects/_template/project.md and the contract docs from PR1-PR4 (cve_authority.*, governance.*, security_inbox.*, forwarders.*, archive_system.*, scope_detection.*). Plus a final scrub of 4 skills for leftover ASF/Vulnogram literals. Byte-equivalent for the airflow-s adopter: every ASF/Airflow/ Vulnogram-specific value either resolves through a config knob whose ASF default matches today's behaviour, OR stays as one named-example aside in generic prose. Per-target lifts: - docs/security/threat-model.md (+107/-77) — Purpose/Scope/ Assumptions reframed from "ASF"/"PMC" to governance-knob terms. STRIDE matrix rows A.6/A.7/C.1-C.4/E.1-E.2 lifted: Vulnogram -> <cve-tool>; security@apache.org -> <security-list>; DRAFT/REVIEW/READY/PUBLIC -> cve_authority.states sequence (allocated -> review-ready -> publish-ready -> public). Mitigations M.10/M.16/M.18/M.19/M.27 + residual risks #3/#8/#10/#11 + re-audit cadence ownership generalised. - docs/security/forwarder-routing-policy.md (+42/-27) — references the optional security-issue-import-via-forwarder sub-skill from PR3 (#387) and the tools/forwarder-relay/README.md contract. Replaces "ASF-security relay" / "security@apache.org" with forwarders.enabled / <security-list> / foundation_security_address. ASF-Airflow shown as a named-example aside per concept. - docs/security/how-to-fix-a-security-issue.md (+20/-8) — "governance-authorised member of the adopting project (per governance.cve_allocation_gate)" replaces "PMC member of apache/airflow"; <cve-tool> + cve_authority.* replaces Vulnogram- specific URLs and state names; archive_system.advisory_publication_signal_url replaces the lists.apache.org users-list URL. - docs/security/new-members-onboarding.md (+26/-13) — onboarding- style register preserved. "PMC members and committers" reframed as "governance body that satisfies governance.cve_allocation_gate"; per-user-config "PMC status" steps reference the governance knob; Vulnogram steps reference <cve-tool> via cve_authority.record_url_template. - Final scrub of 4 skills (+17/-15 net): security-issue-import, security-issue-import-via-forwarder, security-issue-invalidate, security-issue-fix — leftover literal references caught and lifted to roster.bare_name_handles / governance.escalation_contact / forwarders.<adapter>.contact_handle. Aggregate: 8 files, +240/-156 lines. That closes the series. Five PRs (#381, #386, #387, #388, this) transitioned the security skill family from Airflow/ASF-coupled to a generic framework with ASF as the default-configured option. The airflow-s adopter, with the ASF defaults baked into project.md, sees byte-equivalent behaviour throughout. Non-ASF adopters override specific dimensions (CVE authority, mail provider, archive system, governance gate, scope axis) by changing only their <project-config>/ files. Generated-by: Claude Code (Opus 4.7)
potiuk
added a commit
to potiuk/magpie
that referenced
this pull request
Jun 1, 2026
…ity-suite refactor patterns Adds `optimize-skill` (capability:setup) — the refactoring sibling of `write-skill`. It takes an existing framework skill (or sweeps a set) and applies the five restructuring patterns proven on the security suite, as behavior-preserving proposals gated by the validator (green-before / green-after): - split — slim an oversized SKILL.md into linked siblings (the apache#410 pattern; addresses the PRINCIPLES.md P14 cap) - config-lift — move concrete values into <project-config> (apache#386/apache#387/apache#388) - out-of-context — read/PATCH one field without loading the body (apache#412 github-body-field, apache#424 github-rollup) - fetch-upfront — batch per-item round-trips (apache#347) - preflight-classifier — skip obvious no-ops before LLM passes (apache#414/apache#416) SKILL.md is 297 lines; the pass catalogue (smell / exemplar PR / mechanics / behavior-preservation guarantee / validation) lives in the patterns.md sibling. Reads only framework-internal files, so no injection-guard / Privacy-LLM callouts. Ships a step-diagnose eval (5 auto-comparable cases incl. an injection-resistance case) so the skill is not released without an eval (P8). Wires the skill into the capability->skill map and the eval index. Generated-by: Claude Code (Opus 4.8)
potiuk
added a commit
to potiuk/magpie
that referenced
this pull request
Jun 1, 2026
…ity-suite refactor patterns Adds `optimize-skill` (capability:setup) — the refactoring sibling of `write-skill`. It takes an existing framework skill (or sweeps a set) and applies the five restructuring patterns proven on the security suite, as behavior-preserving proposals gated by the validator (green-before / green-after): - split — slim an oversized SKILL.md into linked siblings (the apache#410 pattern; addresses the PRINCIPLES.md P14 cap) - config-lift — move concrete values into <project-config> (apache#386/apache#387/apache#388) - out-of-context — read/PATCH one field without loading the body (apache#412 github-body-field, apache#424 github-rollup) - fetch-upfront — batch per-item round-trips (apache#347) - preflight-classifier — skip obvious no-ops before LLM passes (apache#414/apache#416) SKILL.md is 297 lines; the pass catalogue (smell / exemplar PR / mechanics / behavior-preservation guarantee / validation) lives in the patterns.md sibling. Reads only framework-internal files, so no injection-guard / Privacy-LLM callouts. Ships a step-diagnose eval (5 auto-comparable cases incl. an injection-resistance case) so the skill is not released without an eval (P8). Wires the skill into the capability->skill map and the eval index. Generated-by: Claude Code (Opus 4.8)
potiuk
added a commit
that referenced
this pull request
Jun 1, 2026
…ity-suite refactor patterns (#427) Adds `optimize-skill` (capability:setup) — the refactoring sibling of `write-skill`. It takes an existing framework skill (or sweeps a set) and applies the five restructuring patterns proven on the security suite, as behavior-preserving proposals gated by the validator (green-before / green-after): - split — slim an oversized SKILL.md into linked siblings (the #410 pattern; addresses the PRINCIPLES.md P14 cap) - config-lift — move concrete values into <project-config> (#386/#387/#388) - out-of-context — read/PATCH one field without loading the body (#412 github-body-field, #424 github-rollup) - fetch-upfront — batch per-item round-trips (#347) - preflight-classifier — skip obvious no-ops before LLM passes (#414/#416) SKILL.md is 297 lines; the pass catalogue (smell / exemplar PR / mechanics / behavior-preservation guarantee / validation) lives in the patterns.md sibling. Reads only framework-internal files, so no injection-guard / Privacy-LLM callouts. Ships a step-diagnose eval (5 auto-comparable cases incl. an injection-resistance case) so the skill is not released without an eval (P8). Wires the skill into the capability->skill map and the eval index. Generated-by: Claude Code (Opus 4.8)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Second of 5 PRs converting the security skill family from
Airflow/ASF-coupled to a generic framework with ASF as the
default-configured option.
PR1 (#381)
landed the schema + adapter contracts with ASF defaults. This
PR lifts 6 skills to read those knobs (and the existing sibling
adopter-config files like
scope-labels.md,release-trains.md,fix-workflow.md,milestones.md) instead of inliningASF/Airflow values.
Byte-equivalent for the airflow-s adopter. Every value the
skill currently inlines either becomes a reference to a config
knob whose ASF default matches today's inlined value, OR keeps
the inlined airflow-s value as a single named example in
generic prose.
Per-skill lifts
security-tracker-stats-dashboarddefault-config.yamlannotated with cross-references toscope_detection.labels+tracker.labels; no literal default values changedsecurity-issue-deduplicatesecurity-issue-import-from-md<security-list>placeholder uses, body-field heading map →tracker.body_fields, label list →tracker.labels, scope rule →scope_detection.labelssecurity-issue-fixuv/breeze/prek) →<project-config>/fix-workflow.md; package registries →release_process.artifact_registries;apache/airflow→<upstream>;main→<default-branch>security-issue-triagescope_detection.labels; canned-response examples reframed as airflow-s named examples;@-handle routing → roster referencessecurity-issue-import-from-prscope_detection.labels;Apache Airflow:title-prefix →<vendor>: <product>:derived fromproject.mdAggregate: +280/-129 lines across 7 files.
What is not in this PR
No ASF-default adapter is touched.
tools/vulnogram/,tools/ponymail/,tools/gmail/asf-relay.mdcontinue to bethe only shipping adapters and continue to be referenced
where they always have been. Those rename to
tools/cve-tool-vulnogram/(PR4) and the forwarder-relay /mail-archive sub-tool extracts (PR3) come later.
No skill outside the 6 above is touched. Deep skills
(
security-issue-sync,security-cve-allocate,security-issue-invalidate,security-issue-import) arePR3/PR4 — they need the sub-tool extracts to land first.
No new placeholders are introduced beyond those declared
in PR1's schema. Existing AGENTS.md placeholders
(
<upstream>,<tracker>,<security-list>,<default-branch>, etc.) are used per convention.Coming up
PR3 — forwarder-relay + mail-archive sub-tools.
Renames
tools/ponymail/and the asf-relay shape asASF-default adapters; updates
security-issue-import(drop
ASF-security relayrow from generic body, push intooptional sub-skill),
security-issue-invalidateStep 5d,security-issue-syncStep 2b.PR4 — CVE-authority sub-tool extract (biggest).
Renames
tools/vulnogram/→tools/cve-tool-vulnogram/.Rewrites
security-cve-allocate,security-issue-syncSteps 5b/5c (~600 lines),
security-issue-invalidateStep 0,security-issue-deduplicate,docs/security/process.mdSteps 12-14,
docs/security/roles.mdagainst the<cve-tool>placeholder + tool-agnostic state verbs(
allocated→review-ready→publish-ready→public).PR5 — docs + final scrub.
Test plan
uv run --project tools/skill-and-tool-validator skill-and-tool-validateclean (1 pre-existing soft warning on an unrelated skill).
pytestclean for the validator (218 tests).check-placeholders, etc.).
cross-references and the "airflow-s as named example"
pattern reads well.
existing ASF behaviour without changes
(byte-equivalence invariant — guarded by the fact that PR1's
ASF defaults are byte-equivalent to the previously-hardcoded
values).