@tangle-network/starter-foundry 0.7.3 → 0.13.2

package/registry/families/agent-runtime-photographer-ts/files/methodology/lighting-setup-design.md

@@ -0,0 +1,34 @@
+# Lighting Setup Design
+
+## Purpose
+A structured framework for designing lighting setups for studio or location shoots.
+
+## When to use
+- Planning a studio shoot
+- Solving a lighting challenge
+- When the user asks for lighting advice
+
+## Template
+
+### Subject & Goal
+- **Subject type:** [portrait, product, still life, etc.]
+- **Desired look:** [high key, low key, dramatic, soft, etc.]
+- **Available equipment:** [list lights, modifiers, etc.]
+
+### Light Placement
+- **Key light:** [position, modifier, intensity]
+- **Fill light:** [position, modifier, intensity, ratio to key]
+- **Backlight / Rim light:** [position, modifier, intensity]
+- **Background light:** [if applicable]
+- **Practical lights:** [any ambient or practical sources]
+
+### Light Quality
+- **Hard vs soft:** [modifier choice and distance]
+- **Color temperature:** [kelvin values, gels if any]
+- **Contrast ratio:** [key:fill ratio]
+
+### Trade-offs
+- [What does this setup sacrifice? e.g., loss of texture, increased shadows]
+
+### Alternative Setups
+- [List 1-2 alternative lighting configurations]

package/registry/families/agent-runtime-photographer-ts/files/methodology/post-production-workflow.md

@@ -0,0 +1,32 @@
+# Post-Production Workflow
+
+## Purpose
+A structured framework for post-processing images from import to export.
+
+## When to use
+- Planning an editing session
+- Troubleshooting a specific editing challenge
+- When the user asks for post-production advice
+
+## Template
+
+### Image Context
+- **File type:** [RAW, JPEG, etc.]
+- **Software:** [Lightroom, Capture One, Photoshop, etc.]
+- **Goal:** [color grade, retouch, composite, etc.]
+
+### Workflow Steps
+1. **Import & cull:** [select keepers, rate]
+2. **Basic adjustments:** [exposure, contrast, white balance, highlights, shadows, whites, blacks]
+3. **Tone curve:** [global or channel-specific adjustments]
+4. **Color grading:** [HSL, split toning, calibration]
+5. **Local adjustments:** [gradients, brushes, radial filters]
+6. **Detail:** [sharpening, noise reduction]
+7. **Lens corrections:** [profile corrections, manual adjustments]
+8. **Export:** [format, resolution, color space]
+
+### Trade-offs
+- [What does this workflow sacrifice? e.g., time vs quality, natural vs stylized]
+
+### Alternative Approaches
+- [List 1-2 alternative workflows or techniques]

package/registry/families/agent-runtime-photographer-ts/manifest.json

@@ -0,0 +1,148 @@
+{
+  "id": "agent-runtime-photographer-ts",
+  "description": "Creative photographer agent bundle \u2014 composition coaching, lighting setup design, and post-production workflow guidance. Works alongside a photographer as a thinking partner for shot planning, gear selection, and creative direction. Not a substitute for hands-on experience or professional critique.",
+  "tags": [
+    "agent-runtime",
+    "photographer",
+    "creative-agent",
+    "composition",
+    "lighting",
+    "post-production",
+    "cloudflare-worker"
+  ],
+  "taxonomy": {
+    "language": "markdown",
+    "runtime": "cloudflare-worker",
+    "surface": "agent-runtime"
+  },
+  "includes": [
+    "agent-base:tangle",
+    "agent-base:secure",
+    "agent-output:blocks",
+    "agent-base:memory",
+    "agent-base:scheduler",
+    "agent-base:mcp-registry",
+    "agent-channels:telegram",
+    "agent-channels:discord",
+    "agent-channels:slack",
+    "agent-channels:whatsapp",
+    "agent-channels:imessage",
+    "agent-channels:gmail",
+    "agent-channels:linear"
+  ],
+  "defaults": {
+    "projectType": "agent-runtime-photographer",
+    "agentName": "photographer",
+    "publicVars": [
+      "AGENT_NAME"
+    ],
+    "routes": [
+      {
+        "path": "/api/chat",
+        "auth": "bearer",
+        "callsTools": [
+          "composition-coach",
+          "lighting-setup",
+          "post-production-workflow"
+        ]
+      },
+      {
+        "path": "/api/health",
+        "auth": "none",
+        "callsTools": []
+      }
+    ],
+    "declaredCapabilities": [
+      "composition-coaching",
+      "lighting-setup-design",
+      "post-production-workflow"
+    ],
+    "runtime": {
+      "sandbox": "tangle",
+      "noNewPrivileges": true,
+      "capDrop": "ALL"
+    },
+    "stakes": "low",
+    "advisoryOnly": false,
+    "outboundDomains": [
+      "api.tangle.tools"
+    ],
+    "secrets": [
+      "TANGLE_API_KEY"
+    ]
+  },
+  "files": [
+    {
+      "source": "files/agent.json",
+      "target": "agent.json"
+    },
+    {
+      "source": "files/AGENTS.md",
+      "target": "AGENTS.md"
+    },
+    {
+      "source": "files/methodology/index.json",
+      "target": "methodology/index.json"
+    },
+    {
+      "source": "files/methodology/composition-canvas.md",
+      "target": "methodology/composition-canvas.md"
+    },
+    {
+      "source": "files/methodology/lighting-setup-design.md",
+      "target": "methodology/lighting-setup-design.md"
+    },
+    {
+      "source": "files/methodology/post-production-workflow.md",
+      "target": "methodology/post-production-workflow.md"
+    },
+    {
+      "source": "files/README.md",
+      "target": "README.md"
+    },
+    {
+      "source": "files/TOOLS.md",
+      "target": "TOOLS.md"
+    }
+  ],
+  "contextHints": {
+    "commands": [],
+    "entrypoints": [
+      "AGENTS.md"
+    ],
+    "extensionPoints": [
+      "templates/",
+      "src/lib/tools/"
+    ]
+  },
+  "tieredKeywords": {
+    "tier1": [
+      "photographer-agent",
+      "photographer agent",
+      "composition-coach",
+      "composition coach",
+      "lighting-designer",
+      "lighting designer",
+      "creative-director-agent",
+      "creative director agent"
+    ],
+    "tier2": [
+      "composition",
+      "lighting-setup",
+      "lighting setup",
+      "post-production",
+      "post production",
+      "exposure-triangle",
+      "exposure triangle",
+      "depth-of-field",
+      "depth of field",
+      "color-grading",
+      "color grading"
+    ],
+    "tier3": [],
+    "archetypes": [
+      "agent-in-sandbox",
+      "creative-agent"
+    ]
+  }
+}

package/registry/families/agent-runtime-physics-research-ts/files/AGENTS.md

@@ -0,0 +1,69 @@
+---
+name: physics-research
+role: Physics research assistant — literature search, citation-grounded analysis, and experimental design support
+ domain: physics-research
+allowedDomains:
+  - api.tangle.tools
+  - export.arxiv.org
+  - api.semanticscholar.org
+  - api.openalex.org
+  - api.crossref.org
+allowedEnv:
+  - TANGLE_API_KEY
+advisoryOnly: true
+version: 0.1.0
+---
+
+## Role
+
+You are a physics research assistant. You help researchers, students, and enthusiasts with **literature search**, **citation-grounded analysis**, and **experimental design support**. You are not a substitute for peer review, ethics board approval, or professional judgment. You do not make decisions for the researcher. You do not fabricate data or citations.
+
+## Authoritative skills
+
+When the user's request maps to one of these capabilities, load the corresponding template *before* responding. The templates are the methodology source of truth; trust them over training.
+
+- `literature-search` → `templates/literature-search.md`
+- `citation-analysis` → `templates/citation-analysis.md`
+- `experimental-design` → `templates/experimental-design.md`
+
+## Output blocks
+
+Wrap structured deliverables in parseable blocks the host UI renders distinctly:
+
+- `:::artifact` — literature search results, citation analysis reports, experimental design outlines. Always tag the producing template (e.g. `template: literature-search`).
+- `:::survey` — citation-grounded context produced via the research-corpus tools (arxiv / Semantic Scholar / OpenAlex / Crossref). Inline cite by `[surname, year]`; collect full citations at the bottom of the block. Refuse to fabricate citations.
+- `:::escalation` — emitted whenever a request crosses into territory that requires a real professional (see "Mandatory escalation"). The block names the kind of professional and the question to bring them.
+
+## Mandatory escalation (advisory boundary)
+
+Emit a `:::escalation` block whenever ANY of these fire:
+
+1. **Human subjects research** — IRB approval, consent forms, privacy compliance. → researcher's IRB or ethics board.
+2. **Classified or export-controlled research** — ITAR, EAR, national security restrictions. → researcher's institutional security office.
+3. **Patent or intellectual property advice** — patentability, prior art, licensing. → patent attorney or technology transfer office.
+4. **Medical or clinical advice** — diagnosis, treatment, drug interactions. → licensed medical professional.
+5. **Anything triggering "I should ask my advisor / ethics board / lawyer"** — escalate before advising.
+
+Do not silently rationalize past any of these. State the escalation, name the professional, and offer to help the researcher **prepare** for that conversation (frame the question, list the documents, draft the ask) — preparation is on-scope; the professional opinion itself is not.
+
+## Research-corpus discipline
+
+When the user asks for literature context, related work, benchmark data, or "what does the literature say about X" — **use the research-corpus tools**. Do not hallucinate papers, results, or citations. Search arxiv / Semantic Scholar / OpenAlex / Crossref, cite by `[surname, year]`, and emit the result inside a `:::survey` block.
+
+If you cannot find a citation, say so plainly and downgrade the claim to a hypothesis the user can test — never fabricate.
+
+## What you will NOT do
+
+- Make a decision the researcher is accountable for
+- Replace peer review, ethics board, or professional judgment
+- Fabricate data, citations, or experimental results
+- Give legal, patent, or medical advice (escalate instead)
+- Run a literature search on stale data — re-pull every query
+
+## What you WILL do
+
+- Conduct thorough literature searches using the research-corpus tools
+- Provide citation-grounded analysis and summaries
+- Help design experiments with clear hypotheses, controls, and statistical power
+- Cite, never hallucinate, when a claim requires evidence
+- Pair every escalation-trigger with a concrete handoff: which professional, which document, which question

package/registry/families/agent-runtime-physics-research-ts/files/README.md

@@ -0,0 +1,26 @@
+# Physics Research Assistant Agent
+
+An agent-runtime bundle for physics research assistance: literature search, citation analysis, and experimental design support.
+
+## Features
+
+- **Literature Search**: Query arxiv, Semantic Scholar, OpenAlex, and Crossref for relevant papers.
+- **Citation Analysis**: Analyze citation networks and impact metrics.
+- **Experimental Design**: Help design physics experiments with clear hypotheses and controls.
+
+## Usage
+
+Deploy as a Cloudflare Worker. Set the `TANGLE_API_KEY` environment variable. The agent exposes two routes:
+
+- `POST /api/chat` — Main chat endpoint (requires bearer token)
+- `GET /api/health` — Health check
+
+## Templates
+
+- `templates/literature-search.md` — Systematic literature search protocol
+- `templates/citation-analysis.md` — Citation network analysis protocol
+- `templates/experimental-design.md` — Experimental design protocol
+
+## Disclaimer
+
+This agent is advisory only. It does not replace peer review, ethics board approval, or professional judgment. Always verify results and consult appropriate professionals for high-stakes decisions.

package/registry/families/agent-runtime-physics-research-ts/files/agent.json

@@ -0,0 +1,36 @@
+{
+  "name": "physics-research",
+  "description": "Physics research assistant agent bundle — literature search, citation-grounded analysis, and experimental design support. Composes the agent-runtime substrate with the research-corpus layer (arxiv / Semantic Scholar / OpenAlex / Crossref) so every claim is citation-grounded. Advisory only — never replaces peer review, ethics board, or professional judgment.",
+  "version": "0.1.0",
+  "tags": [
+    "agent-runtime",
+    "physics-research",
+    "literature-search",
+    "citation-grounded",
+    "experimental-design",
+    "cloudflare-worker"
+  ],
+  "prompt": {
+    "systemPromptFile": "AGENTS.md"
+  },
+  "tools": {
+    "filesystem": true,
+    "shell": true,
+    "router": true
+  },
+  "permissions": {
+    "filesystem.write": "ask",
+    "shell.exec": "ask",
+    "network": "ask"
+  },
+  "resources": {
+    "files": [
+      {
+        "source": "templates/"
+      },
+      {
+        "source": "README.md"
+      }
+    ]
+  }
+}

package/registry/families/agent-runtime-physics-research-ts/files/methodology/citation-analysis.md

@@ -0,0 +1,110 @@
+---
+capability: citation-analysis
+status: active
+source: hand-authored, aligned with INSPIRE-HEP and Semantic Scholar citation-graph conventions
+retrieved: 2026-04-26
+---
+
+# Citation Analysis methodology (physics)
+
+Citation analysis maps the structure of a research conversation:
+who cites whom, where consensus is forming, where forks remain, and
+which paper is the currently-cited authoritative measurement or
+calculation. The deliverable is a `:::artifact` with a graph
+description, not a vibes-based assertion of importance.
+
+## When to use
+
+Trigger this template when the user asks:
+
+- "Who has cited <paper>?"
+- "What is the most-cited paper on <topic>?"
+- "Which result do experimentalists / theorists actually use?"
+- "Has <result> been replicated / contradicted?"
+- "What's the lineage of <method/idea>?"
+
+## Method
+
+1. **Identify the seed paper(s).** Start from a specific arXiv ID or
+   DOI. If the user only has a topic, route through
+   `literature-search.md` first to get a seed.
+2. **Pull forward and backward citations.**
+   - Forward (papers that cite the seed): Semantic Scholar
+     `references.cited_by`, INSPIRE-HEP for HEP, OpenAlex for general
+   - Backward (papers the seed cites): the paper's bibliography,
+     OpenAlex `references`
+3. **Tag each citing paper** by its relationship to the seed:
+   - **Builds on**: extends the result in the same direction
+   - **Constrains**: independent measurement bounding the same
+     observable
+   - **Contradicts**: incompatible measurement / calculation
+   - **Reviews**: cites as one of many in a survey
+   - **Methods**: uses the seed's method, not its result
+   - **Tangential**: namedrops the seed without engaging
+4. **Identify the convergence cluster.** Most physics topics have a
+   group of 5–15 papers that everyone in the area cites. Surface that
+   cluster explicitly — it's the "operative consensus."
+5. **Identify forks.** Where two clusters cite different seed papers
+   for the same observable, name the fork: which assumption / data
+   set / regime each cluster takes.
+6. **Highlight contradictions.** Tension > 3σ between two reported
+   measurements is interesting. Name the experiments, the values,
+   the systematics each side flags, and the next decisive measurement.
+7. **Output the graph as a `:::artifact`.**
+
+## Output shape
+
+```
+:::artifact
+template: citation-analysis
+seed: { arxiv: "XXXX.YYYYY", title: "...", year: 2023 }
+forward-citations: 142 (as of 2026-04-26)
+backward-citations: 38
+clusters:
+  - name: "operative consensus"
+    members: ["arXiv:...", "arXiv:...", ...]
+    summary: "..."
+  - name: "fork-A: assume <X>"
+    members: [...]
+  - name: "fork-B: assume <Y>"
+    members: [...]
+contradictions:
+  - { a: "arXiv:1", b: "arXiv:2", observable: "<param>",
+      tension: "3.4σ", reconciliation-attempts: ["arXiv:3"] }
+key-citing-paper: "arXiv:..." (most-cited downstream paper)
+:::
+```
+
+## Citation discipline
+
+Citation counts come from real corpus queries with the date stamped.
+Counts drift; never quote a count without "as of <date>." Refuse to
+invent counts or relationships.
+
+For HEP, INSPIRE-HEP citation counts are the community-standard. For
+broader physics, Semantic Scholar and OpenAlex disagree by
+double-digit percentages on count — report both and note the
+discrepancy when material.
+
+## Common analysis failures
+
+1. **Conflating citation count with importance.** Highly-cited papers
+   are sometimes wrong in ways that took years to detect. A paper
+   widely cited *to be refuted* is still highly cited.
+2. **Ignoring the self-citation network.** Same group citing its own
+   prior work is not independent confirmation.
+3. **Missing the meta-paper.** Physics topics often have a single
+   review or PDG entry that consolidates the field; downstream papers
+   cite the review, not the primary measurements.
+4. **Dropping the date.** Citation graphs update; what was a 3σ
+   tension in 2022 may be reconciled by 2025.
+5. **No-engagement tagging.** Many "tangential" citations are just
+   namedrops. Don't count them as substantive engagement.
+
+## Scope limits
+
+This template is a citation-graph descriptor, not a bibliometric
+ranking. For h-index, journal impact, or career evaluation, route to
+a different tool — these are a different (and more contested)
+question, and the agent should refuse to produce them as authoritative
+evaluations of researchers.

package/registry/families/agent-runtime-physics-research-ts/files/methodology/experimental-design.md

@@ -0,0 +1,129 @@
+---
+capability: experimental-design
+status: active
+source: hand-authored, aligned with PRD/PRL methods conventions and PDG averaging norms
+retrieved: 2026-04-26
+---
+
+# Experimental Design methodology (physics)
+
+The agent helps a researcher draft a defensible measurement or
+calculation. It does not approve the design — that is the
+collaboration's review board, the PI, or peer review. The output is a
+structured `:::artifact` for a methods section, internal note, or
+proposal.
+
+## When to use
+
+Trigger this template when the user asks:
+
+- "How would I measure <observable>?"
+- "What systematic uncertainties matter for <experiment>?"
+- "How precise do I need to be to distinguish <model A> from <model B>?"
+- "How do I propagate uncertainty through <derivation>?"
+- "What controls / null tests should this calculation include?"
+
+If the user has not chosen what they're measuring or computing, route
+to `literature-search.md` first.
+
+## Method
+
+1. **Pin the observable.** State precisely what number is being
+   measured or computed, with units, expected order of magnitude, and
+   the regime of validity. "Cross section" is not enough — at what
+   √s, for what final state, fiducial vs total.
+2. **Pin the discriminating power required.** What is the precision
+   target, and what does it discriminate? "Constrain m_W to ±10 MeV
+   to discriminate between <Model A> and <SM> at >3σ given current
+   electroweak fit."
+3. **Identify the dominant statistical and systematic uncertainties.**
+   Build the full error budget on paper before running anything.
+   - Statistical: scales as 1/√N; sets the integrated luminosity /
+     sample-size requirement
+   - Systematic: each source enumerated — calibration, modeling,
+     background, theory; each gets a budget
+4. **Define controls and null tests.** Every measurement has a "do
+   the same thing on a sample with no signal" or "swap a known
+   ingredient and verify the answer changes correctly." Name them.
+5. **Define blinding strategy.** Hidden-signal-region, scrambled-data,
+   floating-offset; commit to unblinding criteria *before* looking at
+   data. "We unblind once <list of QC checks> all pass" — written
+   down.
+6. **Specify the analysis chain.** From raw data to the reported
+   number: reconstruction, selection, calibration, fit, background
+   subtraction, unfolding (if any), uncertainty propagation. Each step
+   has an input, an output, and a check.
+7. **Specify cross-checks.** Independent analyzers, alternative
+   methods (matrix-element vs neural net, profile likelihood vs
+   sideband), data/MC agreement plots. State the disagreement
+   threshold that triggers re-analysis.
+8. **Pre-register the analysis when applicable.** For collaboration
+   measurements, internal note + analysis plan before unblinding.
+   For independent / preprint work, OSF or arXiv-stamped methodology
+   draft before final results.
+
+## Output shape
+
+```
+:::artifact
+template: experimental-design
+observable: { name: "...", units: "...", regime: "..." }
+discriminating-power:
+  precision-target: "..."
+  models-distinguished: ["...", "..."]
+  significance-target: "3σ"
+error-budget:
+  statistical: "δ_stat ~ 1/√N → need N ≥ ..."
+  systematics:
+    - { source: "calibration", budget: "...", method: "..." }
+    - { source: "modeling", budget: "...", method: "..." }
+controls:
+  null-test: "..."
+  closure-test: "..."
+blinding:
+  strategy: "hidden signal region 0.9 < x < 1.1"
+  unblinding-criteria: ["...", "..."]
+analysis-chain:
+  - step: "reconstruction"
+    input: "raw"
+    output: "..."
+    check: "..."
+cross-checks:
+  - "alternative method: <method-B>"
+  - "second analyzer: <person>"
+:::
+```
+
+## Common design failures
+
+1. **Underestimated systematics.** Assuming all error is statistical
+   gives a misleadingly small budget. Most modern experiments are
+   systematics-dominated.
+2. **No unblinding criteria.** Looking before commitment leads to
+   experimenter bias; pin criteria up front.
+3. **Single analyzer / single method.** Internal cross-checks (two
+   analyzers, two methods) catch ~half of subtle bugs.
+4. **No closure test.** "Run the same pipeline on simulated data
+   where the answer is known" should always succeed before running on
+   data.
+5. **Dropped correlations.** Combining sub-measurements without their
+   covariance gives wrong final uncertainty. Use BLUE or profile
+   likelihood, not naive weighted average, when correlations exist.
+6. **Reporting one-sigma confidence as "evidence."** PDG/HEP
+   convention: 3σ = evidence, 5σ = discovery. Many other fields differ
+   — be explicit about the threshold convention used.
+
+## Refusal triggers
+
+Emit a `:::escalation` block whenever the user requests:
+
+- A study involving **human subjects** (e.g., medical-physics
+  patient data) → IRB / institutional ethics
+- **Classified, ITAR, or export-controlled** work → institutional
+  security office
+- A **patent / IP** opinion on novelty → patent attorney / TTO
+- Any **clinical-physics** dose calculation tied to a real patient →
+  licensed medical physicist + treating clinician
+
+State the escalation, name the body, and offer to help prepare the
+ask (frame the question, draft the outline, list the documents).

package/registry/families/agent-runtime-physics-research-ts/files/methodology/index.json

@@ -0,0 +1,25 @@
+{
+  "entries": [
+    {
+      "id": "literature-search",
+      "capability": "literature-search",
+      "path": "./literature-search.md",
+      "tags": ["methodology", "literature", "arxiv", "inspire-hep", "search"],
+      "version": "0.1.0"
+    },
+    {
+      "id": "citation-analysis",
+      "capability": "citation-analysis",
+      "path": "./citation-analysis.md",
+      "tags": ["methodology", "citation", "graph", "inspire", "openalex"],
+      "version": "0.1.0"
+    },
+    {
+      "id": "experimental-design",
+      "capability": "experimental-design",
+      "path": "./experimental-design.md",
+      "tags": ["methodology", "experiment", "design", "systematics", "blinding"],
+      "version": "0.1.0"
+    }
+  ]
+}