[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7APVA0QNF76SN2TYSN1ZYSZ":3,"skill-same:01M7APVA0QNF76SN2TYSN1ZYSZ":51},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":34,"ownerDescription":35,"updatedAt":36,"install":37,"resources":50},"01M7APVA0QNF76SN2TYSN1ZYSZ","atlas-synthesize",null,"Phase S (Synthesize) — emit the scout-report.md artifact that downstream agents (SPECTRA, APIVR-Δ, human) consume. Hard cap 3000 tokens; every factual clause cited with FINDING-XXX; every recommended action carries a handoff label. Use after Phase A (Abstract) fold is validated and DECISION_TARGET has an answer or its gaps are recorded. Final phase of an ATLAS mission.","Rynaro\u002Fmagicite","Rynaro","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F5271543?v=4","01M70CFYFMB316W23THVQAQ42M","Henrique A. Lavezzo","01M7APN2KCVWBNY727T57KR2ZJ",2,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fatlas\u002Fskills\u002Fsynthesize\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: atlas-synthesize\ndescription: Phase S (Synthesize) — emit the scout-report.md artifact that downstream agents (SPECTRA, APIVR-Δ, human) consume. Hard cap 3000 tokens; every factual clause cited with FINDING-XXX; every recommended action carries a handoff label. Use after Phase A (Abstract) fold is validated and DECISION_TARGET has an answer or its gaps are recorded. Final phase of an ATLAS mission.\nallowed-tools: memex_read view_file\nmetadata:\n  methodology: ATLAS\n  phase: S\n---\n\n# SKILL: Synthesize — scout report emission (Phase S)\n\n## When to use\n\n**Load when:** fold summary is validated and all `DECISION_TARGET` sub-questions\nare either answered or recorded as gaps.\n\n**Unload when:** `scout-report.md` is emitted and handed off.\n\n---\n\n## Contract\n\n| Field | Value |\n|-------|-------|\n| LLM calls permitted | Yes — this is the substantive synthesis edge. |\n| Tool budget | ≤ 15% of mission `max_tool_calls`. Mostly `memex.read` for quoted excerpts. |\n| Output | `scout-report.md` conforming to `templates\u002Fscout-report.md` |\n\n---\n\n## What this phase does\n\nEmit the single artifact downstream agents will consume. Everything\nupstream was preparation; this is the delivery. The report must be:\n\n- **Self-contained.** A SPECTRA or APIVR-Δ agent receiving it should not\n  need to re-explore the codebase.\n- **Decision-ready.** Answers `DECISION_TARGET` explicitly, not implicitly.\n- **Handoff-labeled.** Every recommended action is tagged with the right\n  downstream agent.\n- **Small.** Hard cap 3000 tokens total. If you need more, the mission\n  should have been split.\n\n---\n\n## Structure (mandatory, in order)\n\n### 1. Mission recap (≤150 tokens)\n\nCopy verbatim from `mission.md`:\n\n- `MISSION-ID`, `GOAL`, `DECISION_TARGET`, `SCOPE`.\n\nDo not paraphrase. Downstream agents key off these fields.\n\n### 2. Topology summary (≤500 tokens, ≤10 bullets)\n\nDistilled from `map.md`. Each bullet is one structural fact with at least\none `path` reference. Example:\n\n```\n- HTTP entrypoint: POST \u002Fballots\u002F:id\u002Fcast → BallotsController#cast\n  (config\u002Froutes.rb:42).\n- Single FlowObject handles vote recording: RecordVote\n  (app\u002Fflows\u002Fvote_casting\u002Frecord_vote.rb).\n- Authorization policy: VotingAuthorizer\n  (app\u002Fpolicies\u002Fvoting_authorizer.rb).\n- No background-job path; synchronous only [GAP-003, M].\n- Test coverage: test\u002Fflows\u002Frecord_vote_test.rb (42 cases).\n```\n\n### 3. Answer to DECISION_TARGET (≤800 tokens)\n\nThe substantive deliverable. Every factual clause is followed by a\n`FINDING-XXX` reference. If a question couldn't be answered, say so and\ncite the `GAP-XXX`.\n\nThis section is where the LLM earns its keep. Graph retrieval and AST\nparsing got the facts; synthesis turns them into an answer shaped like what\na senior engineer would write in a PR description or a design doc comment.\n\n**Style rules:**\n\n- One answer per sub-question. No meandering.\n- Confidence tier flows through: M-tier findings warrant \"likely\" language;\n  L-tier findings are flagged as assumptions.\n- No new claims beyond what findings establish. If you notice a gap while\n  writing, stop and decide: spawn follow-up mission, or record in §5.\n\n### 4. Recommended next actions (≤600 tokens)\n\nRanked list. Each item:\n\n```\nR-1 [→ SPECTRA] | priority: high\n    Draft spec for hardening authorization on RecordVote#call to cover\n    the anonymous-ballot case [GAP-003].\n    References: FINDING-017, GAP-003.\n\nR-2 [→ APIVR-Δ] | priority: medium\n    Add a test asserting that RecordVote#call refuses guest users.\n    References: FINDING-017. Estimated timebox: up to 0.5 days.\n\nR-3 [→ human]   | priority: high\n    Decide policy for retry behavior when cast_vote_records write fails\n    mid-request. No existing convention.\n    References: GAP-003, FINDING-021.\n```\n\n**Handoff labels are mechanical:**\n\n- `→ SPECTRA` — needs spec generation before implementation.\n- `→ APIVR-Δ` (or `→ APIVR`) — spec is clear; ready for implementation loop.\n- `→ human` — blocked on a judgment call you cannot make.\n- `→ ATLAS` — deserves a follow-up scout mission (max recursion = 1).\n\n### 5. Risks & gaps (≤300 tokens)\n\nEvery `GAP-XXX` surfaced in Abstract, plus any risk identified during\nsynthesis. Each entry gets a confidence tier and, if possible, a proposed\nmitigation.\n\n### 6. Telemetry (≤100 tokens)\n\n```\nphase   | tokens_in | tokens_out | tool_calls\nA (Assess)      | 412   | 280    | 0\nT (Traverse)    | 2,104 | 620    | 14\nL (Locate)      | 8,730 | 1,840  | 47\nA (Abstract)    | 1,520 | 1,980  | 3\nS (Synthesize)  | 2,100 | 2,450  | 6\nTOTAL           |14,866 | 7,170  | 70\n\nfold_ratio: 0.09\nη (search efficiency): 0.28\n```\n\n---\n\n## Synthesis anti-patterns\n\n| Anti-pattern | Why it's wrong | Fix |\n|--------------|----------------|-----|\n| Restating every finding | Report bloat, obscures the answer | §3 is prose, not a findings dump; the index is already in the fold |\n| Adding new claims (\"also, I noticed...\") | Bypasses evidence anchoring | New observations = new findings = back to Locate |\n| Omitting confidence language | Downstream misreads M as H | \"Likely\", \"not directly verified\", \"inferred from X\" — but only when tier is M or L |\n| Generic next actions | Wastes downstream agent's context | Every recommended action has a concrete anchor and a labeled handoff |\n| Answering questions the mission didn't ask | Scope creep | If it's valuable, propose it as R-N with `→ ATLAS` follow-up tag |\n\n---\n\n## Handoff emission\n\nAfter `scout-report.md` is written, emit exactly one handoff block:\n\n```\n\u003Chandoff>\n  \u003Cprimary_recipient>SPECTRA\u003C\u002Fprimary_recipient>\n  \u003Cfallback_recipient>human\u003C\u002Ffallback_recipient>\n  \u003Creport_path>artifacts\u002FATLAS\u002Fscout-report-MISSION-042.md\u003C\u002Freport_path>\n  \u003Ccritical_gaps>GAP-003\u003C\u002Fcritical_gaps>\n  \u003Copen_questions>\n    - Retry policy on partial CVR write (R-3)\n  \u003C\u002Fopen_questions>\n\u003C\u002Fhandoff>\n```\n\nDownstream tooling parses this block deterministically.\n\n### Envelope sidecar (ECL v2.0)\n\nImmediately after the `\u003Chandoff>` block, also emit a `scout-report.envelope.json`\nfile adjacent to the scout report. This is a terminal Phase-S artefact in the\nsame class as `scout-report.md` — it is produced by the harness as Phase-S\noutput, NOT via a write tool call, preserving the I-1 read-only invariant.\n\nUse `templates\u002Fscout-report.envelope.json` as the fill-in-the-blank skeleton.\nKey fields to populate:\n\n- `artifact.path` — relative path to the `scout-report.md` file (e.g.\n  `artifacts\u002FATLAS\u002Fscout-report-MISSION-042.md`).\n- `artifact.sha256` + `integrity.value` — SHA-256 hex digest of the scout-report\n  file bytes. Compute with `shasum -a 256 \u003Cfile> | awk '{print $1}'` or\n  equivalent. The values MUST match.\n- `artifact.size_bytes` — byte count of the scout-report file\n  (`wc -c \u003C \u003Cfile> | tr -d '[:space:]'`).\n- `from.version` — ATLAS SemVer at the time of emission (e.g. `\"1.13.0\"`).\n- `to.eidolon` \u002F `to.version` — primary handoff recipient (usually `\"spectra\"`).\n  See `SPEC.md §6` for recipient mapping.\n- `performative` — `\"PROPOSE\"` for ATLAS→SPECTRA scout handoffs (ECL v2.0 §2;\n  closed ten-performative set, unchanged from v1.0).\n- `objective` — one sentence ≤240 chars matching the mission GOAL.\n- `trace.host`, `trace.model`, `trace.ts` — fill from the active session context.\n\n**ISE block (ECL v2.0 §6.5, optional but always emitted by ATLAS):**\n\n- `ise.assertion_grade` — always `\"self-attested\"`. Scout-report claims are\n  evidence-anchored (I-7: every claim carries `path:line` + confidence tier)\n  but are ATLAS's own internal validation, not an externally-verified check by\n  another Eidolon — `self-attested` is the accurate grade, not `validated` and\n  not `unverified`.\n- `ise.provenance.methodology_version` — `\"atlas-\u003Cversion>\"` (e.g.\n  `\"atlas-1.13.0\"`), pattern `^[a-z][a-z0-9-]*-\\d+\\.\\d+\\.\\d+$`.\n- `ise.provenance.tool_surface` — optional; the subset of ATLAS's seven\n  read-only tools actually invoked this mission.\n- `ise.provenance.lateral_consults` — optional; record a FORGE consult here\n  if one occurred (`{\"eidolon\": \"forge\", \"performative\": \"PROPOSE\"}`).\n- `ise.receiver_authorization` — always\n  `{\"auto_route\": true, \"auto_merge\": false, \"auto_deploy\": false}`. SPECTRA\n  or APIVR-Δ MAY auto-route the scout report into their own intake without an\n  operator confirm, but MUST NOT auto-merge or auto-deploy anything on the\n  strength of a scout report alone (§6.5.3\u002F§6.5.6 — receivers MUST NOT take an\n  action whose corresponding flag is `false`).\n\nReference contract: `eidolons-ecl\u002Fcontracts\u002Fatlas-to-spectra.yaml` (in the\n`Rynaro\u002Feidolons-ecl` repo) defines the normative `from`\u002F`to`\u002F`edge_origin`\u002F\n`performative` values for the ATLAS→SPECTRA edge.\n\nValidate the emitted envelope against `schemas\u002Fecl-envelope.v2.json` before\nmarking Phase S complete.\n\n**Change-worthy findings:** if this mission surfaced ≥1 finding that asserts a\ndefect, a spec\u002Fimpl drift, or a genuine gap, and the consumer project is\nESL-enabled (`.spectra\u002F` present), load `skills\u002Fesl-hop\u002FSKILL.md` before finalizing\nthis section — it frames the ESL proposal inside the same `scout-report.md` +\nenvelope you are already emitting (no new artefact).\n\n### CRYSTALIUM ingest (memory persistence)\n\nAfter the envelope sidecar is validated, persist the handoff to CRYSTALIUM:\n\n```\nmcp__crystalium__ingest(\n  envelope = \u003Cthe validated scout-report.envelope.json contents>,\n  payload  = \u003Cscout-report.md contents>\n)\n```\n\nThis records the scout-report at T1 with full ECL provenance (`from.eidolon=atlas`\ndrives tier derivation; `integrity.value` is stored as `provenance.content_hash`).\nThe read-only constraint (I-1) applies to the codebase — calling memory MCP tools\nis explicitly allowed and does not violate I-1.\n\n**Direct episodic notes (optional):** For notable mid-cycle observations not worth\na full handoff (e.g. a non-obvious naming convention, a recurring gap pattern),\ncall:\n\n```\nmcp__crystalium__commit(\n  layer      = \"episodic\",\n  payload    = \u003Cobservation>,\n  provenance = { author_agent: \"atlas\", mission_id: \u003CMISSION-ID> }\n)\n```\n\n`author_agent` MUST be `\"atlas\"` on every direct commit.\n\n### Session end\n\nAfter `ingest` completes (or after the handoff block if CRYSTALIUM is absent),\ncall:\n\n```\nmcp__crystalium__session_end()\n```\n\nThis triggers Dream consolidation asynchronously. Call it once per mission\ncompletion.\n\n**Graceful skip:** if `mcp__crystalium__*` tools are unavailable (CRYSTALIUM not\ninstalled), proceed without memory — skip the ingest and session_end calls and\nmark Phase S complete normally. Never hard-fail on absent CRYSTALIUM tools.\n\n---\n\n## Exit gate\n\n- [ ] All six sections present in order.\n- [ ] Total report ≤ 3000 tokens.\n- [ ] Every factual clause in §3 has a FINDING-XXX reference.\n- [ ] Every R-N in §4 has a concrete handoff label.\n- [ ] Every GAP-XXX from the fold appears in §5.\n- [ ] Handoff block emitted and well-formed.\n- [ ] Envelope sidecar emitted, schema-valid against `schemas\u002Fecl-envelope.v2.json`, `integrity.sha256` matches payload.\n- [ ] `ise.assertion_grade: \"self-attested\"` present with `ise.receiver_authorization` set to `{auto_route: true, auto_merge: false, auto_deploy: false}`.\n- [ ] `mcp__crystalium__ingest` called with the envelope + payload (or CRYSTALIUM absent and skip noted).\n- [ ] `mcp__crystalium__session_end` called once (or CRYSTALIUM absent and skip noted).\n- [ ] Memex remains intact; downstream agents can dereference anchors.\n\nIf exit gate fails, the mission did not complete. Report `STATUS: partial`\nin the recap and escalate — do not ship a half-formed report.\n","# SKILL: Synthesize — scout report emission (Phase S)\n\n## When to use\n\n**Load when:** fold summary is validated and all `DECISION_TARGET` sub-questions\nare either answered or recorded as gaps.\n\n**Unload when:** `scout-report.md` is emitted and handed off.\n\n---\n\n## Contract\n\n| Field | Value |\n|-------|-------|\n| LLM calls permitted | Yes — this is the substantive synthesis edge. |\n| Tool budget | ≤ 15% of mission `max_tool_calls`. Mostly `memex.read` for quoted excerpts. |\n| Output | `scout-report.md` conforming to `templates\u002Fscout-report.md` |\n\n---\n\n## What this phase does\n\nEmit the single artifact downstream agents will consume. Everything\nupstream was preparation; this is the delivery. The report must be:\n\n- **Self-contained.** A SPECTRA or APIVR-Δ agent receiving it should not\n  need to re-explore the codebase.\n- **Decision-ready.** Answers `DECISION_TARGET` explicitly, not implicitly.\n- **Handoff-labeled.** Every recommended action is tagged with the right\n  downstream agent.\n- **Small.** Hard cap 3000 tokens total. If you need more, the mission\n  should have been split.\n\n---\n\n## Structure (mandatory, in order)\n\n### 1. Mission recap (≤150 tokens)\n\nCopy verbatim from `mission.md`:\n\n- `MISSION-ID`, `GOAL`, `DECISION_TARGET`, `SCOPE`.\n\nDo not paraphrase. Downstream agents key off these fields.\n\n### 2. Topology summary (≤500 tokens, ≤10 bullets)\n\nDistilled from `map.md`. Each bullet is one structural fact with at least\none `path` reference. Example:\n\n```\n- HTTP entrypoint: POST \u002Fballots\u002F:id\u002Fcast → BallotsController#cast\n  (config\u002Froutes.rb:42).\n- Single FlowObject handles vote recording: RecordVote\n  (app\u002Fflows\u002Fvote_casting\u002Frecord_vote.rb).\n- Authorization policy: VotingAuthorizer\n  (app\u002Fpolicies\u002Fvoting_authorizer.rb).\n- No background-job path; synchronous only [GAP-003, M].\n- Test coverage: test\u002Fflows\u002Frecord_vote_test.rb (42 cases).\n```\n\n### 3. Answer to DECISION_TARGET (≤800 tokens)\n\nThe substantive deliverable. Every factual clause is followed by a\n`FINDING-XXX` reference. If a question couldn't be answered, say so and\ncite the `GAP-XXX`.\n\nThis section is where the LLM earns its keep. Graph retrieval and AST\nparsing got the facts; synthesis turns them into an answer shaped like what\na senior engineer would write in a PR description or a design doc comment.\n\n**Style rules:**\n\n- One answer per sub-question. No meandering.\n- Confidence tier flows through: M-tier findings warrant \"likely\" language;\n  L-tier findings are flagged as assumptions.\n- No new claims beyond what findings establish. If you notice a gap while\n  writing, stop and decide: spawn follow-up mission, or record in §5.\n\n### 4. Recommended next actions (≤600 tokens)\n\nRanked list. Each item:\n\n```\nR-1 [→ SPECTRA] | priority: high\n    Draft spec for hardening authorization on RecordVote#call to cover\n    the anonymous-ballot case [GAP-003].\n    References: FINDING-017, GAP-003.\n\nR-2 [→ APIVR-Δ] | priority: medium\n    Add a test asserting that RecordVote#call refuses guest users.\n    References: FINDING-017. Estimated timebox: up to 0.5 days.\n\nR-3 [→ human]   | priority: high\n    Decide policy for retry behavior when cast_vote_records write fails\n    mid-request. No existing convention.\n    References: GAP-003, FINDING-021.\n```\n\n**Handoff labels are mechanical:**\n\n- `→ SPECTRA` — needs spec generation before implementation.\n- `→ APIVR-Δ` (or `→ APIVR`) — spec is clear; ready for implementation loop.\n- `→ human` — blocked on a judgment call you cannot make.\n- `→ ATLAS` — deserves a follow-up scout mission (max recursion = 1).\n\n### 5. Risks & gaps (≤300 tokens)\n\nEvery `GAP-XXX` surfaced in Abstract, plus any risk identified during\nsynthesis. Each entry gets a confidence tier and, if possible, a proposed\nmitigation.\n\n### 6. Telemetry (≤100 tokens)\n\n```\nphase   | tokens_in | tokens_out | tool_calls\nA (Assess)      | 412   | 280    | 0\nT (Traverse)    | 2,104 | 620    | 14\nL (Locate)      | 8,730 | 1,840  | 47\nA (Abstract)    | 1,520 | 1,980  | 3\nS (Synthesize)  | 2,100 | 2,450  | 6\nTOTAL           |14,866 | 7,170  | 70\n\nfold_ratio: 0.09\nη (search efficiency): 0.28\n```\n\n---\n\n## Synthesis anti-patterns\n\n| Anti-pattern | Why it's wrong | Fix |\n|--------------|----------------|-----|\n| Restating every finding | Report bloat, obscures the answer | §3 is prose, not a findings dump; the index is already in the fold |\n| Adding new claims (\"also, I noticed...\") | Bypasses evidence anchoring | New observations = new findings = back to Locate |\n| Omitting confidence language | Downstream misreads M as H | \"Likely\", \"not directly verified\", \"inferred from X\" — but only when tier is M or L |\n| Generic next actions | Wastes downstream agent's context | Every recommended action has a concrete anchor and a labeled handoff |\n| Answering questions the mission didn't ask | Scope creep | If it's valuable, propose it as R-N with `→ ATLAS` follow-up tag |\n\n---\n\n## Handoff emission\n\nAfter `scout-report.md` is written, emit exactly one handoff block:\n\n```\n\u003Chandoff>\n  \u003Cprimary_recipient>SPECTRA\u003C\u002Fprimary_recipient>\n  \u003Cfallback_recipient>human\u003C\u002Ffallback_recipient>\n  \u003Creport_path>artifacts\u002FATLAS\u002Fscout-report-MISSION-042.md\u003C\u002Freport_path>\n  \u003Ccritical_gaps>GAP-003\u003C\u002Fcritical_gaps>\n  \u003Copen_questions>\n    - Retry policy on partial CVR write (R-3)\n  \u003C\u002Fopen_questions>\n\u003C\u002Fhandoff>\n```\n\nDownstream tooling parses this block deterministically.\n\n### Envelope sidecar (ECL v2.0)\n\nImmediately after the `\u003Chandoff>` block, also emit a `scout-report.envelope.json`\nfile adjacent to the scout report. This is a terminal Phase-S artefact in the\nsame class as `scout-report.md` — it is produced by the harness as Phase-S\noutput, NOT via a write tool call, preserving the I-1 read-only invariant.\n\nUse `templates\u002Fscout-report.envelope.json` as the fill-in-the-blank skeleton.\nKey fields to populate:\n\n- `artifact.path` — relative path to the `scout-report.md` file (e.g.\n  `artifacts\u002FATLAS\u002Fscout-report-MISSION-042.md`).\n- `artifact.sha256` + `integrity.value` — SHA-256 hex digest of the scout-report\n  file bytes. Compute with `shasum -a 256 \u003Cfile> | awk '{print $1}'` or\n  equivalent. The values MUST match.\n- `artifact.size_bytes` — byte count of the scout-report file\n  (`wc -c \u003C \u003Cfile> | tr -d '[:space:]'`).\n- `from.version` — ATLAS SemVer at the time of emission (e.g. `\"1.13.0\"`).\n- `to.eidolon` \u002F `to.version` — primary handoff recipient (usually `\"spectra\"`).\n  See `SPEC.md §6` for recipient mapping.\n- `performative` — `\"PROPOSE\"` for ATLAS→SPECTRA scout handoffs (ECL v2.0 §2;\n  closed ten-performative set, unchanged from v1.0).\n- `objective` — one sentence ≤240 chars matching the mission GOAL.\n- `trace.host`, `trace.model`, `trace.ts` — fill from the active session context.\n\n**ISE block (ECL v2.0 §6.5, optional but always emitted by ATLAS):**\n\n- `ise.assertion_grade` — always `\"self-attested\"`. Scout-report claims are\n  evidence-anchored (I-7: every claim carries `path:line` + confidence tier)\n  but are ATLAS's own internal validation, not an externally-verified check by\n  another Eidolon — `self-attested` is the accurate grade, not `validated` and\n  not `unverified`.\n- `ise.provenance.methodology_version` — `\"atlas-\u003Cversion>\"` (e.g.\n  `\"atlas-1.13.0\"`), pattern `^[a-z][a-z0-9-]*-\\d+\\.\\d+\\.\\d+$`.\n- `ise.provenance.tool_surface` — optional; the subset of ATLAS's seven\n  read-only tools actually invoked this mission.\n- `ise.provenance.lateral_consults` — optional; record a FORGE consult here\n  if one occurred (`{\"eidolon\": \"forge\", \"performative\": \"PROPOSE\"}`).\n- `ise.receiver_authorization` — always\n  `{\"auto_route\": true, \"auto_merge\": false, \"auto_deploy\": false}`. SPECTRA\n  or APIVR-Δ MAY auto-route the scout report into their own intake without an\n  operator confirm, but MUST NOT auto-merge or auto-deploy anything on the\n  strength of a scout report alone (§6.5.3\u002F§6.5.6 — receivers MUST NOT take an\n  action whose corresponding flag is `false`).\n\nReference contract: `eidolons-ecl\u002Fcontracts\u002Fatlas-to-spectra.yaml` (in the\n`Rynaro\u002Feidolons-ecl` repo) defines the normative `from`\u002F`to`\u002F`edge_origin`\u002F\n`performative` values for the ATLAS→SPECTRA edge.\n\nValidate the emitted envelope against `schemas\u002Fecl-envelope.v2.json` before\nmarking Phase S complete.\n\n**Change-worthy findings:** if this mission surfaced ≥1 finding that asserts a\ndefect, a spec\u002Fimpl drift, or a genuine gap, and the consumer project is\nESL-enabled (`.spectra\u002F` present), load `skills\u002Fesl-hop\u002FSKILL.md` before finalizing\nthis section — it frames the ESL proposal inside the same `scout-report.md` +\nenvelope you are already emitting (no new artefact).\n\n### CRYSTALIUM ingest (memory persistence)\n\nAfter the envelope sidecar is validated, persist the handoff to CRYSTALIUM:\n\n```\nmcp__crystalium__ingest(\n  envelope = \u003Cthe validated scout-report.envelope.json contents>,\n  payload  = \u003Cscout-report.md contents>\n)\n```\n\nThis records the scout-report at T1 with full ECL provenance (`from.eidolon=atlas`\ndrives tier derivation; `integrity.value` is stored as `provenance.content_hash`).\nThe read-only constraint (I-1) applies to the codebase — calling memory MCP tools\nis explicitly allowed and does not violate I-1.\n\n**Direct episodic notes (optional):** For notable mid-cycle observations not worth\na full handoff (e.g. a non-obvious naming convention, a recurring gap pattern),\ncall:\n\n```\nmcp__crystalium__commit(\n  layer      = \"episodic\",\n  payload    = \u003Cobservation>,\n  provenance = { author_agent: \"atlas\", mission_id: \u003CMISSION-ID> }\n)\n```\n\n`author_agent` MUST be `\"atlas\"` on every direct commit.\n\n### Session end\n\nAfter `ingest` completes (or after the handoff block if CRYSTALIUM is absent),\ncall:\n\n```\nmcp__crystalium__session_end()\n```\n\nThis triggers Dream consolidation asynchronously. Call it once per mission\ncompletion.\n\n**Graceful skip:** if `mcp__crystalium__*` tools are unavailable (CRYSTALIUM not\ninstalled), proceed without memory — skip the ingest and session_end calls and\nmark Phase S complete normally. Never hard-fail on absent CRYSTALIUM tools.\n\n---\n\n## Exit gate\n\n- [ ] All six sections present in order.\n- [ ] Total report ≤ 3000 tokens.\n- [ ] Every factual clause in §3 has a FINDING-XXX reference.\n- [ ] Every R-N in §4 has a concrete handoff label.\n- [ ] Every GAP-XXX from the fold appears in §5.\n- [ ] Handoff block emitted and well-formed.\n- [ ] Envelope sidecar emitted, schema-valid against `schemas\u002Fecl-envelope.v2.json`, `integrity.sha256` matches payload.\n- [ ] `ise.assertion_grade: \"self-attested\"` present with `ise.receiver_authorization` set to `{auto_route: true, auto_merge: false, auto_deploy: false}`.\n- [ ] `mcp__crystalium__ingest` called with the envelope + payload (or CRYSTALIUM absent and skip noted).\n- [ ] `mcp__crystalium__session_end` called once (or CRYSTALIUM absent and skip noted).\n- [ ] Memex remains intact; downstream agents can dereference anchors.\n\nIf exit gate fails, the mission did not complete. Report `STATUS: partial`\nin the recap and escalate — do not ship a half-formed report.\n",{"name":5,"metadata":24,"description":7,"allowed-tools":27},{"phase":25,"methodology":26},"S","ATLAS","memex_read view_file","main","38c3a82583abf9806de764f89ac0099ad690933a",".eidolons\u002Fatlas\u002Fskills\u002Fsynthesize",0,"https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite","https:\u002F\u002Fgithub.com\u002FRynaro","https:\u002F\u002Fhlavezzo.me","A Code Alchemist with a passion for weaving tales and inscribing transmutation circles in pure code.","2026-10-14 02:35:37",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":49},[39,42,45],{"tool":40,"command":41},"npx","npx skills add Rynaro\u002Fmagicite --skill atlas-synthesize",{"tool":43,"command":44},"bunx","bunx skills add Rynaro\u002Fmagicite --skill atlas-synthesize",{"tool":46,"command":47},"pnpm","pnpm dlx skills add Rynaro\u002Fmagicite --skill atlas-synthesize","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fatlas\u002Fskills\u002Fsynthesize\u002FSKILL.md",true,[],[52,59,66,73,80,87],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M7AQ4GHJQVV8TJBJ8T1V7ZQ9","wine-dxvk-cache-clear","Clear a stale DXVK shader cache that is causing graphical corruption in a\nWine or Proton game. Use when: a game shows corrupted textures or shader\nflicker that started after a driver or DXVK version change.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Ftoy-registry\u002Fskills\u002Fwine-dxvk-cache-clear\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M7AQ4CY1NHSMY1HCZ91387YY","steam-download-region-fix","Fix a Steam client stuck at very low download speed by changing its\ndownload region. Use when: Steam downloads are far below the local\ninternet speed for every game, not just one.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Ftoy-registry\u002Fskills\u002Fsteam-download-region-fix\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M7AQ49P7Y0CDRSYMPVABX80K","proton-battleye-eac-toggle","Toggle BattlEye or Easy Anti-Cheat runtime support for a Steam Proton game.\nUse when: a multiplayer game refuses to launch with an anti-cheat error under Proton.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Ftoy-registry\u002Fskills\u002Fproton-battleye-eac-toggle\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M7AQ45326FAP3X4Q86FK1X4N","vivi-verify-incoming","Load when reading any upstream artefact handed off to Vivi that carries a sibling .envelope.json. BLOCKING per ECL §6.2.2: the orchestrator MUST have verified the envelope's SHA-256 (eidolons verify-envelope --block \u002F eidolons run --verify) and recorded a verify_pass before dispatch. If no verify_pass exists for the message_id, or a verify_fail is present, REFUSE to process the payload and hand back to the orchestrator. Symmetric receiver gate — every Eidolon enforces it identically.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fvivi\u002Fskills\u002Fverify-incoming\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M7AQ402861P0P3C2TEGFGMA1","vivi-parallel-tracks","Load ONLY under TRANCE authorization when the Plan phase yields N independent, non-overlapping implementation tracks (disjoint file sets). Operationalizes the TRANCE G4 form — parallel multi-track implementation in mandatory git-worktree isolation, max 5 clean-context tracks, a per-track verifier cascade, a non-fungible per-track ≤3 reflection budget, explicit stop conditions, and a single-threaded merge\u002Faggregation step. Absent TRANCE gating, Vivi runs exactly as today: single-track A→P→I→V→Δ\u002FR.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fvivi\u002Fskills\u002Fparallel-tracks\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M7AQ3RH9BZF9ZMTC7VG5P7PB","vivi-methodology","Full Vivi cycle reference — A(nalyze) → P(lan) → I(mplement) → V(erify) → Δ(Delta)\u002FR(eflect). Use when the task is a non-trivial feature implementation in a brownfield codebase and you need the complete methodology reference: complexity routing, evidence-grounded planning rules, test-anchoring requirements, and failure-escalation thresholds.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fvivi\u002Fskills\u002Fmethodology\u002FSKILL.md",[],[]]