|
| 1 | +# Security Posture Plan |
| 2 | + |
| 3 | +## Purpose |
| 4 | + |
| 5 | +This plan explains how the Agentic Enterprise Readiness Profile handles security for an independent OSS profile repository. |
| 6 | + |
| 7 | +It is written for maintainers, external reviewers, and AAIF/LF-style community reviewers who need to understand the security boundary before trying the profile against a real repository. |
| 8 | + |
| 9 | +This document is not a certification, audit report, AAIF endorsement, Linux Foundation endorsement, or production-readiness claim. |
| 10 | + |
| 11 | +## Current Security Scope |
| 12 | + |
| 13 | +The repository currently contains: |
| 14 | + |
| 15 | +- Markdown documentation |
| 16 | +- JSON schema |
| 17 | +- JSON examples |
| 18 | +- boundary-model examples |
| 19 | +- issue templates |
| 20 | +- a lightweight local validation script |
| 21 | + |
| 22 | +The repository does not currently operate: |
| 23 | + |
| 24 | +- a hosted service |
| 25 | +- a database |
| 26 | +- an authentication system |
| 27 | +- a payment flow |
| 28 | +- a model provider integration |
| 29 | +- a queue, worker, or background runtime |
| 30 | +- a production data store |
| 31 | + |
| 32 | +Security risk is therefore concentrated in contributor behavior, documentation hygiene, schema integrity, release discipline, and supply-chain hygiene. |
| 33 | + |
| 34 | +## Reporting Path |
| 35 | + |
| 36 | +Security reporting follows `SECURITY.md`. |
| 37 | + |
| 38 | +Use GitHub issues for non-sensitive security concerns such as missing guardrails, unclear secret-safety guidance, broken validation assumptions, or overly broad claims. |
| 39 | + |
| 40 | +Do not open a public issue that includes: |
| 41 | + |
| 42 | +- credentials |
| 43 | +- customer data |
| 44 | +- private vulnerability details |
| 45 | +- screenshots containing tokens |
| 46 | +- proprietary internal architecture |
| 47 | +- exploit instructions that materially increase risk |
| 48 | + |
| 49 | +If sensitive detail is required, contact the maintainer privately first and only publish a sanitized public follow-up after the sensitive material is removed. |
| 50 | + |
| 51 | +## Secret Safety Rules |
| 52 | + |
| 53 | +Contributors must not commit real secrets, private customer evidence, private policy documents, or screenshots containing tokens. |
| 54 | + |
| 55 | +Examples and templates must use: |
| 56 | + |
| 57 | +- placeholder values |
| 58 | +- environment-variable names |
| 59 | +- secret-reference names |
| 60 | +- redacted excerpts |
| 61 | +- synthetic data |
| 62 | + |
| 63 | +The profile should describe required secret boundaries without asking contributors to expose the actual secret values used by their own systems. |
| 64 | + |
| 65 | +## Supply-Chain Posture |
| 66 | + |
| 67 | +The repository is intentionally small and has no runtime service dependencies. |
| 68 | + |
| 69 | +Current supply-chain controls: |
| 70 | + |
| 71 | +- dependency and license posture is documented in `docs/DEPENDENCY_LICENSE_INVENTORY.md` |
| 72 | +- validation runs through the repository's GitHub Actions workflow |
| 73 | +- examples are checked against `schema/agentic-readiness.schema.json` |
| 74 | +- release notes state what changed and which checks ran |
| 75 | +- no vendored third-party code is expected in docs, examples, or templates |
| 76 | + |
| 77 | +Before a release, maintainers should recheck: |
| 78 | + |
| 79 | +- `npm test` |
| 80 | +- `git diff --check` |
| 81 | +- package metadata parsing |
| 82 | +- changed documentation for accidental secrets |
| 83 | +- changed examples against the JSON schema |
| 84 | + |
| 85 | +## OpenSSF-Style Plan |
| 86 | + |
| 87 | +This project should adopt external security hygiene signals in stages instead of claiming maturity too early. |
| 88 | + |
| 89 | +Planned checks: |
| 90 | + |
| 91 | +| Stage | Action | Evidence target | |
| 92 | +| --- | --- | --- | |
| 93 | +| 1 | Keep `SECURITY.md`, this posture plan, dependency inventory, release checklist, and validation workflow current. | Public docs and passing validation workflow. | |
| 94 | +| 2 | Enable OpenSSF Scorecard or an equivalent public security-health workflow once the repository shape stabilizes. | Public workflow result or issue explaining blocked checks. | |
| 95 | +| 3 | Review OpenSSF Best Practices criteria and document which items apply to a documentation/profile repository. | Public gap issue or checklist. | |
| 96 | +| 4 | Consider branch protection, signed releases, and stricter workflow permissions if external contributors or adopters increase. | Repository settings evidence where public, or maintainer note where settings are private. | |
| 97 | + |
| 98 | +Until these checks are actually enabled and passing, the project must describe them as planned, not complete. |
| 99 | + |
| 100 | +## Release Security Gates |
| 101 | + |
| 102 | +Every release should answer: |
| 103 | + |
| 104 | +- Did validation pass? |
| 105 | +- Did docs or examples introduce any real secrets? |
| 106 | +- Did dependency or license posture change? |
| 107 | +- Did the release checklist run? |
| 108 | +- Did the release notes avoid endorsement or production-readiness claims? |
| 109 | +- Did any security-relevant issue remain open, and is it documented? |
| 110 | + |
| 111 | +Security-sensitive releases should prefer smaller changes and clearer evidence over large bundled edits. |
| 112 | + |
| 113 | +## Issue And PR Hygiene |
| 114 | + |
| 115 | +Maintainers should keep public review focused on concrete artifacts. |
| 116 | + |
| 117 | +Good issues and PRs: |
| 118 | + |
| 119 | +- identify one repository boundary, schema field, validator behavior, or documentation gap |
| 120 | +- avoid private screenshots or internal customer names |
| 121 | +- include a local proof path |
| 122 | +- state whether the evidence is static inspection, local validation, or live runtime proof |
| 123 | +- avoid promotional language |
| 124 | + |
| 125 | +Poor issues and PRs: |
| 126 | + |
| 127 | +- ask maintainers to review private data in public |
| 128 | +- claim AAIF, Linux Foundation, enterprise, or production status without evidence |
| 129 | +- add paid-service defaults without OSS alternatives |
| 130 | +- add broad marketing copy without improving repository safety or usability |
| 131 | + |
| 132 | +## AI-Assisted Contribution Boundary |
| 133 | + |
| 134 | +AI-assisted contributions are allowed when they are reviewed by a human maintainer before merge. |
| 135 | + |
| 136 | +AI tools must not be used to: |
| 137 | + |
| 138 | +- paste private secrets into prompts or issues |
| 139 | +- infer unsupported production-readiness claims |
| 140 | +- auto-post outreach at scale |
| 141 | +- open noisy PRs across many repositories |
| 142 | +- replace maintainer judgment on security-sensitive changes |
| 143 | + |
| 144 | +Contributors should disclose AI assistance when it materially shaped a contribution and should keep the proof path reproducible by humans. |
| 145 | + |
| 146 | +## Artifact Integrity |
| 147 | + |
| 148 | +Release notes should link to: |
| 149 | + |
| 150 | +- the release tag |
| 151 | +- the validation workflow run |
| 152 | +- the changed public docs |
| 153 | +- the issue or PR when the release responds to external feedback |
| 154 | + |
| 155 | +This keeps the project reviewable without relying on private chat logs or unpublished claims. |
| 156 | + |
| 157 | +## Recognition Boundary |
| 158 | + |
| 159 | +This posture supports public review and responsible OSS contribution. It does not create: |
| 160 | + |
| 161 | +- AAIF affiliation |
| 162 | +- Linux Foundation affiliation |
| 163 | +- OpenSSF certification |
| 164 | +- SOC 2, ISO, or enterprise security certification |
| 165 | +- production-readiness proof |
| 166 | + |
| 167 | +Those claims require separate formal processes and independent evidence. |
0 commit comments